diff --git a/apps/presentation/dashboard/src/data/fixtures/edgebench-ad-20261023-history.json b/apps/presentation/dashboard/src/data/fixtures/edgebench-ad-20261023-history.json new file mode 100644 index 0000000000..e411881ddd --- /dev/null +++ b/apps/presentation/dashboard/src/data/fixtures/edgebench-ad-20261023-history.json @@ -0,0 +1,9447 @@ +{ + "description": "Complete captured Explore display state from a real EdgeBench run. Original wording, event and result identities, recording times, measurements and references are unchanged.", + "source_sha256": "23a55562f8ad9b808a5633c479451b78b13ad5370a35190d04fab407e3824a72", + "source_bytes": 396701, + "goal_id": "ad-explore-test-20261023-18h03", + "display_name": "Ad placement · 2026-10-23", + "provenance": { + "case_id": "ad_placement_optimization", + "source_run_id": "ad-explore-test-20261023-18h03", + "study_id": "szh-explore-oldnew-schema-ad-20261023-03", + "selection": "Record-rich complete collected Explore logs selected after inspecting newer local captures: the 26 October Ad and Schema logs contain only two and four records. This Ad history exercises repeated result versions; this Schema history exercises inherited records across three UTC dates. Selection is for reading complexity, not score. No events are sampled.", + "capture_kind": "terminal_collected_display_state", + "exported_at": "2026-11-04T07:12:13.320011+00:00", + "claim_scope": "matched_study", + "score_countable": false, + "runner_status": "runner_invalid", + "limitations": [ + "The complete saved Explore log is not the complete solver trajectory or a coverage claim for every attempt.", + "Recorded confirmed labels are source-authored; negative and unknown results remain in their original prose.", + "No explicit parent or relationship edges were recorded. Shared wording, time adjacency and version updates do not establish causal relationships.", + "Historical event times and originating agents are retained, including records inherited by the actual destination run. Recording intervals are not active solver time.", + "Native and local measurement scopes remain distinct. Cancelled or invalid runs do not qualify a scientific comparison." + ], + "redactions": [], + "boundary": "Display-state events only. No authority or credential store, raw session transcript, hidden evaluator, workspace or model weights are included." + }, + "source_counts": { + "event_count": 255, + "event_kinds": { + "node": 65, + "finding": 190 + }, + "statuses": { + "resolved": 40, + "confirmed": 187, + "tentative": 3, + "exploring": 25 + }, + "first_recorded_at": "2026-10-22T17:31:31Z", + "last_recorded_at": "2026-10-23T07:35:25+00:00", + "edge_count": 0, + "parent_records": 0 + }, + "events": [ + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "7e3984cc5e26cec5", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T17:31:31Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Combined local aspect moves and three starts improved mean score over v1 by 876830 points on 20 training seeds and 832427 on 10 held-out seeds. All 34 final full-budget layouts passed strict geometry checks. Native generalization is unmeasured; factors are not isolated. Continue native evaluation and controlled optimization.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "2457c75c8d4e701f", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v1-train/scores.tsv", + "research/results/v1-holdout/scores.tsv" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:7bf7283ae1ee60439cc283fe62708ce9529002503cbe8120163b30401c78a921", + "comparison_scope": "ad-placement:local:train:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v1-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:44053a6bd4fa5b44f4036a9610e0a50370d5399f8426862cb4eb00ae1aa2fbeb", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 988584723.25 + }, + { + "candidate_ref": "sha256:7bf7283ae1ee60439cc283fe62708ce9529002503cbe8120163b30401c78a921", + "comparison_scope": "ad-placement:local:holdout:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v1-holdout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:428db1da3cc2092283b86cdfb9d30874954eae07107037291e8d3a0f54193404", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 987821782.8 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T17:32:42Z", + "result_id": "finding_6c27f6a65ed7", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Reference source 7bf7283a: full-budget means 988584723 on seeds 0-19 and 987821783 on held-out seeds 1000-1009. Local proxy results only; native score unmeasured.", + "title": "Single-start v1 local reference at a 4.7-second budget" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "f0e34968f9ec1f9c", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v3-train/scores.tsv", + "research/results/v3-holdout/scores.tsv", + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:a8f6e8aabf66ea5f9ed79ce5ad82ad40fc10e0581ee8d5c19249346c29599ef2", + "comparison_scope": "ad-placement:local:train:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v3-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:44053a6bd4fa5b44f4036a9610e0a50370d5399f8426862cb4eb00ae1aa2fbeb", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 989461553.05 + }, + { + "candidate_ref": "sha256:a8f6e8aabf66ea5f9ed79ce5ad82ad40fc10e0581ee8d5c19249346c29599ef2", + "comparison_scope": "ad-placement:local:holdout:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v3-holdout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:428db1da3cc2092283b86cdfb9d30874954eae07107037291e8d3a0f54193404", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 988654209.3 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T17:32:53Z", + "result_id": "finding_cfa1269c056f", + "schema_version": "loopx_explore_result_event_v0", + "status": "tentative", + "summary": "Full-budget v3 mean: 989461553 train, 988654209 holdout; paired wins 14/20 and 7/10 against v1. Short 0.7-second screens did not show gains from local aspect moves or lookahead cuts; three starts were supported only at the full budget. v5 is untested. All inputs, exact sources, outputs and timing data are retained. Next discriminating check: native evaluation of source a8f6e8aa, then controlled held-out comparisons.", + "title": "v3 is the best locally supported candidate; native score remains unmeasured" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "555d3fc57d584df6", + "event_kind": "node", + "evidence_refs": [ + "research/native/submission-1.json", + "research/native/result-1-poll2.txt" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T17:43:36Z", + "result_id": "node_15f56f47d4eb", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Submission 6f37be5d9bbb, source a8f6e8aa, passed 50/50 native cases and scored 49443802408. Per-case scores and resource measurements were not exposed. The native validity question for this snapshot is answered; optimization and evaluating newer local candidates remain open.", + "title": "Does validated v3 satisfy native constraints, and what score does its submitted snapshot achieve?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "ace58845d9ba88c9", + "event_kind": "finding", + "evidence_refs": [ + "research/native/result-1-poll2.txt", + "research/native/submission-1.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:a8f6e8aabf66ea5f9ed79ce5ad82ad40fc10e0581ee8d5c19249346c29599ef2", + "comparison_scope": "ad-placement:native:current-task:fixed50", + "evaluation_ref": "native-protocol:README-sha256:0051d2a6e88e299efc0c0b0f1a7e8a68b3060c855b55ffce8c6cfa601d4b30b2", + "evidence_ref": "research/native/result-1-poll2.txt", + "higher_is_better": true, + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "metric": "total_native_score", + "unit": "points_out_of_50e9", + "value": 49443802408 + } + ], + "node_id": "node_15f56f47d4eb", + "recorded_at": "2026-10-22T17:43:45Z", + "result_id": "finding_749f262ac696", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Accepted round agent-1 / 6f37be5d9bbb evaluated retained source a8f6e8aa. Result: 50/50 passed, 100% pass rate, total score 49443802408. This result certifies that snapshot only. No per-case scores, elapsed times or memory values were public. Native comparison scope follows the fixed-50 task protocol; server implementation revision is not exposed. Further optimization remains authorized and useful.", + "title": "v3 passed all 50 native cases with total score 49443802408" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "677ec8a83987cd21", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v5-train/scores.tsv", + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:9aaa8c51bd6cbefab56295a663dce1f60b25bbaae02aa319a06e94f7e7fbec60", + "comparison_scope": "ad-placement:local:train:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v5-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:44053a6bd4fa5b44f4036a9610e0a50370d5399f8426862cb4eb00ae1aa2fbeb", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 988417533.65 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T17:46:23Z", + "result_id": "finding_1a3515accd9f", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Source 9aaa8c51 isolated lookahead cuts and neighbor regrowth within v3. Mean 988417533.65, delta -1044019.40 against v3, with 5/20 wins. All layouts passed strict geometry checks; one end-to-end launch exceeded five seconds despite internal timer 4.7024. Not adopted. This is a scoped negative result, not a proof that every lookahead design is worse.", + "title": "Full-budget v5 lookahead regressed on the fixed training suite" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "29cb324953a4a855", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v6-train/scores.tsv", + "research/results/v6-holdout/scores.tsv", + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:b2534666dfb39a25f91645212d263b3f9cda28a6fee12f168094d9b7cec9c8b0", + "comparison_scope": "ad-placement:local:train:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v6-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:44053a6bd4fa5b44f4036a9610e0a50370d5399f8426862cb4eb00ae1aa2fbeb", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 989858143.8 + }, + { + "candidate_ref": "sha256:b2534666dfb39a25f91645212d263b3f9cda28a6fee12f168094d9b7cec9c8b0", + "comparison_scope": "ad-placement:local:holdout:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v6-holdout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:428db1da3cc2092283b86cdfb9d30874954eae07107037291e8d3a0f54193404", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 989146278.3 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T17:46:34Z", + "result_id": "finding_2259fcd8f0e7", + "schema_version": "loopx_explore_result_event_v0", + "status": "tentative", + "summary": "Source b2534666 avoids growth scans that cannot extend an edge. Gains vs v3: 396590.75 train (13/20 wins), 492069 validation (6/10). All 34 full-budget layouts and four sanitizer edge checks passed. The diagonal stress score regressed to 848613749 from 975403644. V6 remains a retained alternative; v3 stays in solution.cpp because its native score is verified. Next compare natively after cooldown and investigate stress-case variance.", + "title": "V6 early exits improved generated-case means but need native and robustness confirmation" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "9985926d0b01e882", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T17:46:50Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Updated comparison: v6 has higher generated-case means than v3 on train and validation, but regresses on a diagonal stress input. v5 lookahead regressed. V3 remains the native incumbent; v6 is saved for native comparison and robustness investigation. These local findings do not establish distribution-independent gains.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "e8fdb7cd99695ed7", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v3-validation-fresh/scores.tsv", + "research/results/v3-edge/scores.tsv" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:a8f6e8aabf66ea5f9ed79ce5ad82ad40fc10e0581ee8d5c19249346c29599ef2", + "comparison_scope": "ad-placement:local:validation-fresh:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v3-validation-fresh/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:f8a63b9298b14142ff1b7535751a0195d4954c0411f2cb9274fc18bdbba1f1ef", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 988624296.2 + }, + { + "candidate_ref": "sha256:a8f6e8aabf66ea5f9ed79ce5ad82ad40fc10e0581ee8d5c19249346c29599ef2", + "comparison_scope": "ad-placement:local:diagonal:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v3-edge/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:5ee395fca17a103e8e0c55337b7962770ab0a1d8fba6758ca3f8e6b9690f88c5", + "metric": "diagonal_stress_score", + "unit": "points_out_of_1e9", + "value": 975403644 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T18:45:24Z", + "result_id": "finding_35189314e56f", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "The retained native incumbent scored mean 988624296.2 on fresh seeds 2000-2009. Existing diagonal stress reference: 975403644. Fresh inputs were generated before the new candidate was selected; source and resource conditions are explicit in the measurement records. Local proxy and stress results remain separate from native scoring.", + "title": "V3 reference on newly reserved validation seeds and diagonal stress" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "38b4bdae43eefeff", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v6-validation-fresh/scores.tsv", + "research/results/v6-edge/scores.tsv", + "research/fixed-work/comparison.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:b2534666dfb39a25f91645212d263b3f9cda28a6fee12f168094d9b7cec9c8b0", + "comparison_scope": "ad-placement:local:validation-fresh:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v6-validation-fresh/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:f8a63b9298b14142ff1b7535751a0195d4954c0411f2cb9274fc18bdbba1f1ef", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 988718466.6 + }, + { + "candidate_ref": "sha256:b2534666dfb39a25f91645212d263b3f9cda28a6fee12f168094d9b7cec9c8b0", + "comparison_scope": "ad-placement:local:diagonal:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v6-edge/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:5ee395fca17a103e8e0c55337b7962770ab0a1d8fba6758ca3f8e6b9690f88c5", + "metric": "diagonal_stress_score", + "unit": "points_out_of_1e9", + "value": 848613749 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T18:45:38Z", + "result_id": "finding_c8b255047328", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "V6 scored mean 988718466.6 on reserved seeds 2000-2009, a 94170.4 gain over v3 in this run. Its previous diagonal result was 848613749. A separate fixed-iteration diagnostic produced identical v3/v6 outputs on all tested paths; wall-clock phase scheduling still changes search outcomes. Native v6 remains unmeasured.", + "title": "V6 early-exit reference on fresh validation" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "37be86674b705fe6", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v7-train/scores.tsv", + "research/results/v7-validation-fresh/scores.tsv", + "research/results/v7-edge/scores.tsv", + "research/experiment-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:d2c393df46c1a78bb470c767bcfca931f23613c1f75deb0fbd083506d1a87fd0", + "comparison_scope": "ad-placement:local:validation-fresh:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v7-validation-fresh/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:f8a63b9298b14142ff1b7535751a0195d4954c0411f2cb9274fc18bdbba1f1ef", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 989621640.4 + }, + { + "candidate_ref": "sha256:d2c393df46c1a78bb470c767bcfca931f23613c1f75deb0fbd083506d1a87fd0", + "comparison_scope": "ad-placement:local:diagonal:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v7-edge/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:5ee395fca17a103e8e0c55337b7962770ab0a1d8fba6758ca3f8e6b9690f88c5", + "metric": "diagonal_stress_score", + "unit": "points_out_of_1e9", + "value": 999852912 + }, + { + "candidate_ref": "sha256:d2c393df46c1a78bb470c767bcfca931f23613c1f75deb0fbd083506d1a87fd0", + "comparison_scope": "ad-placement:local:train:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v7-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:44053a6bd4fa5b44f4036a9610e0a50370d5399f8426862cb4eb00ae1aa2fbeb", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 989939059.9 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T18:45:52Z", + "result_id": "finding_3a3dae84b786", + "schema_version": "loopx_explore_result_event_v0", + "status": "tentative", + "summary": "Source d2c393df adds 1% proposals that reset and regrow 2-3 spatial neighbors. Mean 989939059.9 train; fresh mean 989621640.4, 903173.8 above v6 with 7/10 wins, and 997344.2 above v3 with 5/10 wins. Diagonal score 999852912. All 34 full-budget layouts passed strict geometry checks; four sanitizer edge checks passed. V7 is saved for native comparison; v3 remains the submitted incumbent. No native gain is claimed.", + "title": "V7 group rebuilding is the next native comparison candidate after fresh validation" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "4c1093c35d71529e", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v8-train/scores.tsv", + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:1cfd0acb8393592ae6dec89161ec61b6a6ee7262610812c8b1696117ce30e328", + "comparison_scope": "ad-placement:local:train:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v8-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:44053a6bd4fa5b44f4036a9610e0a50370d5399f8426862cb4eb00ae1aa2fbeb", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 989925482.45 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T18:46:03Z", + "result_id": "finding_88ad0ea40129", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Source 1cfd0acb reduces group proposals from 1% to 0.1%. Mean training score 989925482.45, 13577.45 below v7; 11/20 individual cases beat v7, so this small difference does not prove a general frequency effect. Strict geometry and sanitizer edge checks passed. V7 was selected by the higher training mean before fresh validation; v8 was not tested on fresh seeds and was not adopted.", + "title": "V8 rare group rebuilds did not establish a training advantage over v7" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "5d14649ee82e7f51", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md", + "research/experiment-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T18:49:12Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "V7 adds rare 2-3-ad group reconstruction and is the next native candidate after selection on training and fresh seeds 2000-2009. It averaged 989621640.4 on fresh cases and reached 999852912 on diagonal stress, with all 34 full-budget geometry checks passing. V8 was nearly tied on training and untested on fresh seeds. Native v7 result is pending; retain v3 fallback.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "95a06e350ea09856", + "event_kind": "node", + "evidence_refs": [ + "research/native/submission-1.json", + "research/native/submission-2.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T18:51:56Z", + "result_id": "node_9a8e7c2231a3", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Comparable fixed-50 native results: v3/source a8f6e8aa scored 49443802408; v7/source d2c393df scored 49483706996, both passing 50/50. Retain v7 as the measured incumbent. The gain is 39904588; continued optimization remains open, with future submissions respecting acceptance cooldown and exact source/result identities.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "70dbcf59a4b54d73", + "event_kind": "finding", + "evidence_refs": [ + "research/native/submission-1.json", + "research/native/result-1-poll2.txt" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:a8f6e8aabf66ea5f9ed79ce5ad82ad40fc10e0581ee8d5c19249346c29599ef2", + "comparison_scope": "ad-placement:native:current-task:fixed50", + "evaluation_ref": "native-protocol:README-sha256:0051d2a6e88e299efc0c0b0f1a7e8a68b3060c855b55ffce8c6cfa601d4b30b2", + "evidence_ref": "research/native/result-1-poll2.txt", + "higher_is_better": true, + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "metric": "total_native_score", + "unit": "points_out_of_50e9", + "value": 49443802408 + } + ], + "node_id": "node_9a8e7c2231a3", + "recorded_at": "2026-10-22T18:52:09Z", + "result_id": "finding_62c2dd34ceab", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Reference result 6f37be5d9bbb / agent-1 evaluated source a8f6e8aa. The fixed-50 task score is 49443802408 and every case passed. This existing result is copied into the native comparison question with its original provenance; it is not a new evaluation.", + "title": "Native v3 incumbent reference: 49443802408, 50/50 passed" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "3350d55f379134a9", + "event_kind": "finding", + "evidence_refs": [ + "research/native/submission-2.json", + "research/native/result-2-poll1.txt" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:d2c393df46c1a78bb470c767bcfca931f23613c1f75deb0fbd083506d1a87fd0", + "comparison_scope": "ad-placement:native:current-task:fixed50", + "evaluation_ref": "native-protocol:README-sha256:0051d2a6e88e299efc0c0b0f1a7e8a68b3060c855b55ffce8c6cfa601d4b30b2", + "evidence_ref": "research/native/result-2-poll1.txt", + "higher_is_better": true, + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "metric": "total_native_score", + "unit": "points_out_of_50e9", + "value": 49483706996 + } + ], + "node_id": "node_9a8e7c2231a3", + "recorded_at": "2026-10-22T18:52:22Z", + "result_id": "finding_b1f0029ed764", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Submission 4e0d46e8bc2b / agent-2 evaluated exact source d2c393df. Native total 49483706996, 50/50 passed, gain 39904588 over v3. solution.cpp is byte-identical to the accepted v7 snapshot. Retain v7. Local evidence supports this route, but the native comparison does not isolate early exits from group reconstruction or certify global optimality. Continue controlled optimization under cooldown.", + "title": "Native v7 improved total score by 39904588 and remains valid on all 50 cases" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "b99b432f41d6a957", + "event_kind": "finding", + "evidence_refs": [ + "research/restart-v9-train-comparison.json", + "research/results/v9-train/scores.tsv", + "research/results/v9-train/strict.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:c5e493e8443ba6e9d083a8b96d84d116374af8966a72a333af7a9d01067787ba", + "comparison_scope": "ad-placement:local:train:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v9-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:44053a6bd4fa5b44f4036a9610e0a50370d5399f8426862cb4eb00ae1aa2fbeb", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990863796.2 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T19:06:18Z", + "result_id": "finding_9386174e1359", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "At the unchanged 4.70-second O2 budget, v9 changes only STARTS from3 to2. On seeds0..19 it scored990863796.2, beating saved v7 by924736.3 with13/20 wins. All20 outputs satisfy strict geometry. This training screen selects v9 for a fresh paired comparison; it does not establish a native improvement.", + "title": "V9 two restarts improved training mean by 924736 points; fresh confirmation pending" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "d5d69a7d6161ad43", + "event_kind": "finding", + "evidence_refs": [ + "research/restart-v10-train-comparison.json", + "research/results/v10-train/scores.tsv", + "research/results/v10-train/strict.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:53145e074bedbacf359acb6ad04055e5fae2e79e92e6ca811104a25936150a91", + "comparison_scope": "ad-placement:local:train:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v10-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:44053a6bd4fa5b44f4036a9610e0a50370d5399f8426862cb4eb00ae1aa2fbeb", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990167888.25 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T19:06:19Z", + "result_id": "finding_a2c2eb035f43", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "At4.70 seconds, v10 changes only STARTS from3 to5. Mean990167888.25 exceeds saved v7 by228828.35 but trails v9 by695907.95 on the same20 training seeds. All20 outputs strictly valid. Prefer v9 for the planned held-out probe; preserve v10 as a measured alternative, not a refuted route.", + "title": "V10 five restarts scored below the two-restart training candidate" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "cc560b74974dcfdd", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v7-restart-heldout/scores.tsv", + "research/restart-heldout-input-manifest.json", + "research/results/v7-restart-heldout/strict.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:d2c393df46c1a78bb470c767bcfca931f23613c1f75deb0fbd083506d1a87fd0", + "comparison_scope": "ad-placement:local:restart-heldout:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v7-restart-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:b95e6adfc683b5911b78fee343d037a8f0a5d9b3ffc96a13477762d5c59a09e9", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990006864.2 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T19:08:00Z", + "result_id": "finding_3b86bb60ab7e", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Seeds3000..3019 were generated before restart selection. V7 and selected v9 ran serially with alternating execution order at4.70 seconds, O2. V7 mean990006864.2; all20 strict geometry checks passed. Preserve this same-input baseline separately from training and native results.", + "title": "Fresh restart comparison: native incumbent v7 reference mean990006864.2" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "4a724a7efeaee937", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v9-restart-heldout/scores.tsv", + "research/restart-v9-heldout-comparison.json", + "research/results/v9-restart-heldout/strict.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:c5e493e8443ba6e9d083a8b96d84d116374af8966a72a333af7a9d01067787ba", + "comparison_scope": "ad-placement:local:restart-heldout:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v9-restart-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:b95e6adfc683b5911b78fee343d037a8f0a5d9b3ffc96a13477762d5c59a09e9", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990642573.95 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T19:08:01Z", + "result_id": "finding_33cbd6331b57", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Selected from training before seeing seeds3000..3019, v9 scored990642573.95 versus paired v7 mean990006864.2 (+635709.75), winning12/20. Both passed20 strict geometry checks, max local elapsed4.734s. The n-stratified gains reversed from training, so no adaptive n rule is justified. V9 is supported for eligible native evaluation; this is not native improvement or Goal completion.", + "title": "V9 two restarts gained635710 mean points on20 fresh paired cases" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "4273cafc42c22a1c", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v11-train/scores.tsv", + "research/contact-v11-train-comparison.json", + "research/results/v11-train/strict.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:3a45b48923beaa44af3c314be4f7dc0f403b56f73cc3709e1f231e639c5fc6eb", + "comparison_scope": "ad-placement:local:train:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v11-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:44053a6bd4fa5b44f4036a9610e0a50370d5399f8426862cb4eb00ae1aa2fbeb", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990145262.35 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T19:14:15Z", + "result_id": "finding_ccc42af5ee31", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "V11 changes only group-neighbor selection from anchor distance to current rectangle-gap distance. Training mean990145262.35 is206202.45 above saved v7 with10/20 wins, but718533.85 below v9. All20 outputs strictly valid. No fresh or native validation: preserve the moderate result and prefer v9 with fresh evidence.", + "title": "V11 rectangle-gap group selection trailed v9 on training" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "c86e210fc629017a", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v13-train/scores.tsv", + "research/restart-v13-train-comparison.json", + "research/results/v13-train/strict.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:0b0a4941bdf0c1c2b0c907cb0a6618e42669b1d8144a3f6498c2b1bc6595060a", + "comparison_scope": "ad-placement:local:train:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v13-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:44053a6bd4fa5b44f4036a9610e0a50370d5399f8426862cb4eb00ae1aa2fbeb", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990232614.7 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T19:14:16Z", + "result_id": "finding_0034ef37c18b", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "V13 changes default STARTS from3 to1. Training mean990232614.7 beats saved v7 by293554.8 with10/20 wins but trails v9 by631181.5. All20 outputs strictly valid. The1/2/3/5-start screen selects2 for native follow-up. Single-start generalization remains unmeasured; no adaptive n rule is inferred.", + "title": "V13 single restart did not continue the two-restart training gain" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "b61f07305a196d2b", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v7-shortcut-heldout/scores.tsv", + "research/shortcut-heldout-input-manifest.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:d2c393df46c1a78bb470c767bcfca931f23613c1f75deb0fbd083506d1a87fd0", + "comparison_scope": "ad-placement:local:shortcut-heldout:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v7-shortcut-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:4988721b602de7e0fb53d2bd7edecbf22edf4e6c9a9062a55bb131938a259349", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990493509.4 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T19:19:31Z", + "result_id": "finding_db54c3617b10", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Seeds4000..4019 were generated before the three-way shortcut probe. V7/v9/v14 ran serially with rotating and reversed order, unchanged4.70-second O2 budget. V7 mean990493509.4; all20 strict geometry checks passed. Same-scope reference for v9/v14, separate from the first fresh restart suite and native results.", + "title": "Second fresh suite: v7 native-incumbent reference mean990493509.4" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "aade05c041a43127", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v9-shortcut-heldout/scores.tsv", + "research/shortcut-v9-heldout-comparison.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:c5e493e8443ba6e9d083a8b96d84d116374af8966a72a333af7a9d01067787ba", + "comparison_scope": "ad-placement:local:shortcut-heldout:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v9-shortcut-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:4988721b602de7e0fb53d2bd7edecbf22edf4e6c9a9062a55bb131938a259349", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990752680.4 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T19:19:33Z", + "result_id": "finding_410cda6e573c", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "On seeds4000..4019 v9 scored990752680.4 versus v7 mean990493509.4 (+259171,12/20 wins), all strict-valid. Together with the earlier fresh20-case result, two restarts gained447440.375 mean points over v7 across40 fresh cases,24/40 wins. This supports native testing, not an adaptive n rule or native improvement.", + "title": "V9 gained259171 mean points on a second20-case fresh suite" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "6a9a2345ffd06ee2", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md", + "research/next-native-candidate.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T19:19:35Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 local tests on generated inputs. V9 improved two fresh suites; v14 led the latest20-case suite by85690 points over v9. Native confirmation pending; v7 remains native incumbent.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "ba9b70109f8b0e7e", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md", + "research/next-native-candidate.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T19:22:58+00:00", + "result_id": "node_e388daf41666", + "run_id": "explore-result:7aa010a6135263e8445a4db7f9bc061685a86f69d66e17c94482a6ee64df7c2b", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 local tests on generated inputs. V9 improved two fresh suites; v14 led the latest20-case suite by85690 points over v9. Native confirmation pending; v7 remains native incumbent.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "1050139d20a19a3e", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-v14-shortcut-heldout" + ], + "execution_observation": { + "candidate_sha256": "da276cc6e410db24217d35dd2311243390564df36af97392a7379306bfacaff0", + "change_reason": null, + "elapsed_seconds": 94.34657, + "evidence_refs": [ + "evidence:ad-v14-shortcut-heldout" + ], + "executed_method_id": "ad-placement.roundrobin.20x4p70s.v1", + "exit_code": 0, + "input_revision": "sha256:4988721b602de7e0fb53d2bd7edecbf22edf4e6c9a9062a55bb131938a259349", + "limit_seconds": 100, + "planned_method_id": "ad-placement.roundrobin.20x4p70s.v1", + "result_candidate_sha256": "da276cc6e410db24217d35dd2311243390564df36af97392a7379306bfacaff0", + "result_ref": "evidence:ad-v14-shortcut-heldout", + "schema_version": "explore_execution_observation_v0", + "search_iterations": 159708180, + "stop_reason": "20 serial searches completed; each used a4.70-second wall budget within the5-second per-case limit." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:da276cc6e410db24217d35dd2311243390564df36af97392a7379306bfacaff0", + "comparison_scope": "ad-placement:local:shortcut-heldout:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "evidence:ad-v14-shortcut-heldout", + "higher_is_better": true, + "input_revision": "sha256:4988721b602de7e0fb53d2bd7edecbf22edf4e6c9a9062a55bb131938a259349", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990838370.5 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T19:22:58+00:00", + "result_id": "finding_7aa010a6135263e8445a4db7f9bc061685a86f69d66e17c94482a6ee64df7c2b", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:4988721b602de7e0fb53d2bd7edecbf22edf4e6c9a9062a55bb131938a259349 Applicability: 4.70-second O2 local tests on generated inputs. V9 improved two fresh suites; v14 led the latest20-case suite by85690 points over v9. Native confirmation pending; v7 remains native incumbent. Observation: V14 averaged990838370.5 on seeds4000..4019: +85690.1 vs v9 and+344861.1 vs v7. All20 strict checks passed; maximum elapsed4.72715s. V14 remains unsubmitted. Interpretation: Stage v14 for eligible native evaluation, preserving v7 as native incumbent. Two restarts gained447440 mean points over v7 across40 fresh cases. The shortcut preserved24 tested fixed-work paths and reduced internal time by9% in one trial. Its extra timed-score gain is small; diagonal stress regressed to917166105 versus earlier v7 999852912. This does not establish native improvement or Goal completion.", + "tags": [ + "writeback-result" + ], + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "b960eb95b235db80", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v14-push-train/scores.tsv", + "research/results/v14-push-train/strict.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:da276cc6e410db24217d35dd2311243390564df36af97392a7379306bfacaff0", + "comparison_scope": "ad-placement:local:train:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v14-push-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:44053a6bd4fa5b44f4036a9610e0a50370d5399f8426862cb4eb00ae1aa2fbeb", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990140113.8 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T19:34:10Z", + "result_id": "finding_63fc86f67b37", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "V14 reran at4.70 seconds against chain-push v15 on seeds0..19 in alternating serial order. Mean990140113.8; all20 outputs strictly valid. This is below the saved v9 training mean990863796.2, while previous fresh v14-v9 comparison was slightly positive. Timing variance and sample selection remain limits; do not hide the training weakness.", + "title": "V14 paired training reference scored990140113.8" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "2431323602e1c892", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v15-train/scores.tsv", + "research/push-v15-train-comparison.json", + "research/push-throughput.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:17f49caae79bbe8bdd149bab50379eb7f6f184c3ce6121247f69e93acac03b22", + "comparison_scope": "ad-placement:local:train:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v15-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:44053a6bd4fa5b44f4036a9610e0a50370d5399f8426862cb4eb00ae1aa2fbeb", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 989420252.25 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T19:34:12Z", + "result_id": "finding_b5bed896dfbf", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Compound directional pushes shift collision chains while preserving orthogonal spans and accept/reject transactionally. At5% proposal frequency, v15 mean989420252.25 trailed paired v14 by719861.55 with8/20 wins; iteration count fell only2.6%. All20 full outputs and24 sanitizer cases were valid. Do not adopt this frequency. A0.625% variant is a separate planned probe, not evidence that this route generally fails.", + "title": "V15 chain pushes at5% frequency lost719862 training mean points" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "f3f4204fbef21100", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v16-train/scores.tsv", + "research/reheat-v16-train-comparison.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:5d29f433a392b813b7cff3952dcc6b22ac80eb2307c67bafd8650804bcc99d95", + "comparison_scope": "ad-placement:local:train:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v16-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:44053a6bd4fa5b44f4036a9610e0a50370d5399f8426862cb4eb00ae1aa2fbeb", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990156616.45 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T19:40:52Z", + "result_id": "finding_2292211ee86c", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "V16 resumes the saved global best for epoch2 instead of reinitializing, preserving v14 temperature and step schedules. At4.70 seconds its mean990156616.45 exceeded the preceding v14 mean by16502.65, with9/20 wins. All20 outputs strictly valid. This does not establish a useful gain; v18 had the highest training mean and was selected for the reserved fresh suite.", + "title": "V16 best-layout reheating nearly tied v14 on training" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "7be9f3f37d0e3551", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v17-train/scores.tsv", + "research/reheat-v17-train-comparison.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:6952ff33d5e4aa766a31e2ba28a09abb6022ed124c1f31189a871e3dc3ea8125", + "comparison_scope": "ad-placement:local:train:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v17-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:44053a6bd4fa5b44f4036a9610e0a50370d5399f8426862cb4eb00ae1aa2fbeb", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 989606221.5 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T19:40:54Z", + "result_id": "finding_32e59a933249", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "V17 resumes the saved best and lowers epoch2 initial temperature from0.025 to0.005 while retaining final7.5e-7. Mean989606221.5 trailed preceding v14 by533892.3,8/20 wins. All20 outputs strictly valid. Preserve this scoped negative result; do not adopt the cooler reheating schedule or infer a general impossibility.", + "title": "V17 cooler best-layout reheating regressed training mean by533892" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "5769b9e3bd4b4ed5", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v18-train/scores.tsv", + "research/push-v18-train-comparison.json", + "research/wake5-selection-plan.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:5ee4b5e0e199fd605916fb5a3c3962d21045b919c80ea4087d6ff497a0d12c4e", + "comparison_scope": "ad-placement:local:train:gxx-O2:4.70s:v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v18-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:44053a6bd4fa5b44f4036a9610e0a50370d5399f8426862cb4eb00ae1aa2fbeb", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990504797.05 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T19:40:55Z", + "result_id": "finding_abf53afe1d34", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "V18 lowers directional chain-push proposals from5% to0.625%, keeping other v14 settings. Training mean990504797.05 beats preceding v14 by364683.25,12/20 wins, with all20 outputs strictly valid. It led v15/v16/v17 and was selected before inspecting seeds5000..5019. This is a selection result; fresh and native confirmation remain necessary.", + "title": "V18 rare chain pushes led the training screen by364683 over v14" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "13ddafdd2a1ec9fd", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v14-wake5-heldout/scores.tsv", + "research/wake5-heldout-input-manifest.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:da276cc6e410db24217d35dd2311243390564df36af97392a7379306bfacaff0", + "comparison_scope": "ad-placement:local:wake5-heldout:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v14-wake5-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:d0b741f0d6456ddfd8b220fef644e1a0c7353744f563cc4e314943580cdaadd9", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990459345.3 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T19:48:15Z", + "result_id": "finding_c1d76931fc40", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Seeds5000..5019 were reserved before selection. V14/v18/v19 ran serially with rotating and reversed order at4.70 seconds, O2. V14 mean990459345.3; all20 strict geometry checks passed. This same-input reference is separate from prior fresh suites and native results.", + "title": "Wake5 fresh suite: v14 reference mean990459345.3" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "3a0f926eb45a7561", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v18-wake5-heldout/scores.tsv", + "research/push-v18-heldout-comparison.json", + "research/v18-heldout-bootstrap.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:5ee4b5e0e199fd605916fb5a3c3962d21045b919c80ea4087d6ff497a0d12c4e", + "comparison_scope": "ad-placement:local:wake5-heldout:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v18-wake5-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:d0b741f0d6456ddfd8b220fef644e1a0c7353744f563cc4e314943580cdaadd9", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990554779.7 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T19:48:16Z", + "result_id": "finding_4d1c902c9da3", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Selected before fresh evaluation, v18 mean990554779.7 exceeds paired v14 by95434.4,13/20 wins. All20 strict checks passed, max4.73322s. The exploratory paired bootstrap95% interval [-675806.35,765156.45] crosses zero and excludes within-case timing variance. Combined with the positive training screen this supports a native trial, not a native gain or a conclusive causal effect.", + "title": "V18 rare chain pushes gained95434 fresh mean points over v14" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "67b958c5c8e2e984", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v19-wake5-heldout/scores.tsv", + "research/polish-fixed-layout-comparison.json", + "research/polish-v19-heldout-comparison.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:367ca611e1ffe79df867766edeff38cf4cd8547501c4b9f20670658499989c4f", + "comparison_scope": "ad-placement:local:wake5-heldout:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v19-wake5-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:d0b741f0d6456ddfd8b220fef644e1a0c7353744f563cc4e314943580cdaadd9", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990504380.35 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T19:48:17Z", + "result_id": "finding_add027073488", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "The standalone final-growth probe gained only4 total tester points on44 saved v14 layouts, with zero regressions and per-ad nondecrease checks. Timed v19 mean990504380.35 on seeds5000..5019 beats v14 by45035.05 but wins only7/20; all outputs valid. Do not attribute this timed difference to greedy completion or treat it as a meaningful quality gain. V18 led this fresh suite.", + "title": "V19 final growth offered negligible saved-layout gains" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "a7c6753959bf838b", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md", + "research/wake5-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T19:48:19Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 local generated tests. V18 rare chain pushes led training and fresh seeds5000-5019; fresh gain95434 points over v14 is uncertain. Reheating and final growth gave no established benefit.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "0e29aeff150291ce", + "event_kind": "node", + "evidence_refs": [ + "research/native/submission-3.json", + "research/native/result-v18-poll3.txt" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T19:57:24Z", + "result_id": "node_9a8e7c2231a3", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Comparable native fixed50 results: v3 49443802408, v7 49483706996, v18 49510341240; all50/50 passed. V18 is the measured incumbent; continued optimization remains open.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "0ade3e128d2903c0", + "event_kind": "node", + "evidence_refs": [ + "research/native/submission-3.json", + "research/native/result-v18-poll3.txt" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T19:59:06+00:00", + "result_id": "node_9a8e7c2231a3", + "run_id": "explore-result:404b8579fbfcee85929cbfa8969fd3df68960e64092c5e008fd7e5f5f2332fbd", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Comparable native fixed50 results: v3 49443802408, v7 49483706996, v18 49510341240; all50/50 passed. V18 is the measured incumbent; continued optimization remains open.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "026f56804c11909f", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-v18-native" + ], + "execution_observation": { + "candidate_sha256": "5ee4b5e0e199fd605916fb5a3c3962d21045b919c80ea4087d6ff497a0d12c4e", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "evidence:ad-v18-native" + ], + "executed_method_id": "sforge.native.fixed50.v1", + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "limit_seconds": null, + "planned_method_id": "sforge.native.fixed50.v1", + "result_candidate_sha256": "5ee4b5e0e199fd605916fb5a3c3962d21045b919c80ea4087d6ff497a0d12c4e", + "result_ref": "evidence:ad-v18-native", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Server completed50/50 cases; per-case times, memory and search iterations were not exposed." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:5ee4b5e0e199fd605916fb5a3c3962d21045b919c80ea4087d6ff497a0d12c4e", + "comparison_scope": "ad-placement:native:current-task:fixed50", + "evaluation_ref": "native-protocol:README-sha256:0051d2a6e88e299efc0c0b0f1a7e8a68b3060c855b55ffce8c6cfa601d4b30b2", + "evidence_ref": "evidence:ad-v18-native", + "higher_is_better": true, + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "metric": "total_native_score", + "unit": "points_out_of_50e9", + "value": 49510341240 + } + ], + "node_id": "node_9a8e7c2231a3", + "recorded_at": "2026-10-22T19:59:06+00:00", + "result_id": "finding_404b8579fbfcee85929cbfa8969fd3df68960e64092c5e008fd7e5f5f2332fbd", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50 Applicability: Comparable native fixed50 results: v3 49443802408, v7 49483706996, v18 49510341240; all50/50 passed. V18 is the measured incumbent; continued optimization remains open. Observation: Accepted v18/source5ee4b5e0, submissionc9348702537c, passed50/50 and scored49510341240: +26634244 versus native v7. Retain the exact accepted snapshot in solution.cpp. Interpretation: V18 is the highest-scoring tested native candidate. Relative to v7 it combines two starts, unchanged-rectangle early returns and rare chain pushes; native gain does not isolate contributions. Local frequent-push and cooler-reheat regressions and negligible final-growth evidence remain preserved. Continue P1 with partial-area aspect proposals; no global optimum or Goal completion is claimed.", + "tags": [ + "writeback-result" + ], + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "7982ef9f24dd806a", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v18-area-train/scores.tsv", + "research/wake6-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:5ee4b5e0e199fd605916fb5a3c3962d21045b919c80ea4087d6ff497a0d12c4e", + "comparison_scope": "ad-placement:local:area-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v18-area-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:44053a6bd4fa5b44f4036a9610e0a50370d5399f8426862cb4eb00ae1aa2fbeb", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990101541.35 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T20:18:41Z", + "result_id": "finding_2fb70924b1b5", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Matched4.70-second baseline averaged990101541.35 on seeds0..19. All20 outputs passed tester and strict geometry.", + "title": "V18 baseline for area-proposal training screen" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "41cd2c588fe212cc", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v20-train/scores.tsv", + "research/wake6-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:53bd2da89b16c626130a16fc03e2574843d8b49c0782df071e9f7560b9a64108", + "comparison_scope": "ad-placement:local:area-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v20-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:44053a6bd4fa5b44f4036a9610e0a50370d5399f8426862cb4eb00ae1aa2fbeb", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991392053.85 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T20:18:43Z", + "result_id": "finding_8ec5a38d6bf3", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "V20 averaged991392053.85: +1290512.50 versus matched v18,13/20 wins. All20 strict checks passed. Current/full area mixture lost to v21 on this screen; preserve as an alternative, not a refuted route.", + "title": "V20 mixed-area proposals improved training mean" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "8276d7ee26b535b2", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v21-train/scores.tsv", + "research/wake6-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:6b580fbd4974af01ea009b62925a510c6c15dabc74476b9f007ef1bce00dfafd", + "comparison_scope": "ad-placement:local:area-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v21-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:44053a6bd4fa5b44f4036a9610e0a50370d5399f8426862cb4eb00ae1aa2fbeb", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992611783.25 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T20:18:43Z", + "result_id": "finding_f58fbf342513", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "V21 averaged992611783.25: +2510241.90 versus matched v18,19/20 wins. Selected before inspecting reserved seeds6000..6019. All20 strict checks passed; native feedback remains unmeasured.", + "title": "V21 current-area proposals led training screen" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "179e5157f8636e99", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v22-train/scores.tsv", + "research/wake6-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:6ec3ca7ede526e5efb71c1fd973a4f29913c5d17561fbdaf02aa740651452d52", + "comparison_scope": "ad-placement:local:area-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v22-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:44053a6bd4fa5b44f4036a9610e0a50370d5399f8426862cb4eb00ae1aa2fbeb", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991739655.95 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T20:18:44Z", + "result_id": "finding_f6acb6f608ba", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "V22 averaged991739655.95: +1638114.60 versus matched v18,16/20 wins. All20 strict checks passed. It lost to v21 on this training screen, without establishing a general disadvantage.", + "title": "V22 slightly enlarged-area proposals improved training mean" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "02bd3da581df059e", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v18-wake6-heldout/scores.tsv", + "research/wake6-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:5ee4b5e0e199fd605916fb5a3c3962d21045b919c80ea4087d6ff497a0d12c4e", + "comparison_scope": "ad-placement:local:wake6-heldout:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v18-wake6-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:dd0d0f5cbed12daeec4855880b44e876f23b56c28797a570b1d6df17a9137d9e", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 989190064.7 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T20:18:45Z", + "result_id": "finding_5db7c0b57a3f", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Matched4.70-second baseline averaged989190064.70 on reserved seeds6000..6019. All20 outputs passed tester and strict geometry. Separate scope from prior held-out cohorts.", + "title": "V18 baseline for fresh current-area proposal comparison" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "e22f9d00ce6e499a", + "event_kind": "node", + "evidence_refs": [ + "research/wake6-summary.json", + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T20:18:55Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 generated tests. V21 current-area aspect proposals led training and gained1445708 fresh mean points over v18 on seeds6000-6019,15/20 wins. Native confirmation is pending.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "60bb1b14e80947ee", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v18-area-train/scores.tsv", + "research/area-train-input-manifest.json", + "research/wake6-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:5ee4b5e0e199fd605916fb5a3c3962d21045b919c80ea4087d6ff497a0d12c4e", + "comparison_scope": "ad-placement:local:area-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v18-area-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990101541.35 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T20:23:53Z", + "result_id": "finding_2fb70924b1b5", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Matched4.70-second baseline averaged990101541.35 on seeds0..19. All20 outputs passed tester and strict geometry. Training input identifier corrected from a direct byte manifest; scores are unchanged.", + "title": "V18 baseline for area-proposal training screen" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "8917675ebed80bd7", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v20-train/scores.tsv", + "research/area-train-input-manifest.json", + "research/wake6-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:53bd2da89b16c626130a16fc03e2574843d8b49c0782df071e9f7560b9a64108", + "comparison_scope": "ad-placement:local:area-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v20-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991392053.85 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T20:23:54Z", + "result_id": "finding_8ec5a38d6bf3", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "V20 averaged991392053.85: +1290512.50 versus matched v18,13/20 wins. All20 strict checks passed. Current/full area mixture lost to v21 on this screen; preserve as an alternative, not a refuted route. Training input identifier corrected from a direct byte manifest; scores are unchanged.", + "title": "V20 mixed-area proposals improved training mean" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "f098440dafb6fb9d", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v21-train/scores.tsv", + "research/area-train-input-manifest.json", + "research/wake6-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:6b580fbd4974af01ea009b62925a510c6c15dabc74476b9f007ef1bce00dfafd", + "comparison_scope": "ad-placement:local:area-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v21-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992611783.25 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T20:23:55Z", + "result_id": "finding_f58fbf342513", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "V21 averaged992611783.25: +2510241.90 versus matched v18,19/20 wins. Selected before inspecting reserved seeds6000..6019. All20 strict checks passed; native feedback remains unmeasured. Training input identifier corrected from a direct byte manifest; scores are unchanged.", + "title": "V21 current-area proposals led training screen" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "f13adb3fb42dc800", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v22-train/scores.tsv", + "research/area-train-input-manifest.json", + "research/wake6-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:6ec3ca7ede526e5efb71c1fd973a4f29913c5d17561fbdaf02aa740651452d52", + "comparison_scope": "ad-placement:local:area-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v22-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991739655.95 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T20:23:56Z", + "result_id": "finding_f6acb6f608ba", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "V22 averaged991739655.95: +1638114.60 versus matched v18,16/20 wins. All20 strict checks passed. It lost to v21 on this training screen, without establishing a general disadvantage. Training input identifier corrected from a direct byte manifest; scores are unchanged.", + "title": "V22 slightly enlarged-area proposals improved training mean" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "acf859df589d0a5a", + "event_kind": "node", + "evidence_refs": [ + "research/wake6-summary.json", + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T20:24:35+00:00", + "result_id": "node_e388daf41666", + "run_id": "explore-result:e0d7cc73772c8372da9ce96d344654e6a9981b1ade72f95205888e442fda4b53", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 generated tests. V21 current-area aspect proposals led training and gained1445708 fresh mean points over v18 on seeds6000-6019,15/20 wins. Native confirmation is pending.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "d70a0bf06a17f469", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-v21-wake6-heldout" + ], + "execution_observation": { + "candidate_sha256": "6b580fbd4974af01ea009b62925a510c6c15dabc74476b9f007ef1bce00dfafd", + "change_reason": null, + "elapsed_seconds": 94.36267000000001, + "evidence_refs": [ + "evidence:ad-v21-wake6-heldout" + ], + "executed_method_id": "ad-placement.paired.20x4p70s.v1", + "exit_code": 0, + "input_revision": "sha256:dd0d0f5cbed12daeec4855880b44e876f23b56c28797a570b1d6df17a9137d9e", + "limit_seconds": 100, + "planned_method_id": "ad-placement.paired.20x4p70s.v1", + "result_candidate_sha256": "6b580fbd4974af01ea009b62925a510c6c15dabc74476b9f007ef1bce00dfafd", + "result_ref": "evidence:ad-v21-wake6-heldout", + "schema_version": "explore_execution_observation_v0", + "search_iterations": 172052244, + "stop_reason": "20 serial searches completed; each had4.70-second search budget under5-second case limit." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:6b580fbd4974af01ea009b62925a510c6c15dabc74476b9f007ef1bce00dfafd", + "comparison_scope": "ad-placement:local:wake6-heldout:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "evidence:ad-v21-wake6-heldout", + "higher_is_better": true, + "input_revision": "sha256:dd0d0f5cbed12daeec4855880b44e876f23b56c28797a570b1d6df17a9137d9e", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990635773.1 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T20:24:35+00:00", + "result_id": "finding_e0d7cc73772c8372da9ce96d344654e6a9981b1ade72f95205888e442fda4b53", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:dd0d0f5cbed12daeec4855880b44e876f23b56c28797a570b1d6df17a9137d9e Applicability: 4.70-second O2 generated tests. V21 current-area aspect proposals led training and gained1445708 fresh mean points over v18 on seeds6000-6019,15/20 wins. Native confirmation is pending. Observation: V21 averaged990635773.10 on seeds6000..6019: +1445708.40 versus matched v18,15/20 wins. All20 strict checks passed; maximum elapsed4.72795s. Source6b580fbd remains unsubmitted. Interpretation: V21 was selected on training before fresh inspection and is the supported next native candidate. Its gain does not establish native generalization. V20/v22 also improved training but lost to v21. Diagonal stress regressed to935396798 versus prior v18 989856126. Preserve native v18 and all snapshots; read eligibility before the source-matched comparison. The20-case execution resources below are batch totals, not per-case times.", + "tags": [ + "writeback-result" + ], + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "890e525a1b71ec4a", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v21-shrink-train/scores.tsv", + "research/wake7-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:6b580fbd4974af01ea009b62925a510c6c15dabc74476b9f007ef1bce00dfafd", + "comparison_scope": "ad-placement:local:shrink-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v21-shrink-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991331147.1 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T20:40:12Z", + "result_id": "finding_c718e37d2287", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Matched4.70-second baseline averaged991331147.10 on seeds0..19. All20 tester and strict geometry checks passed. Baseline differs from earlier timed trials; use within-screen deltas.", + "title": "V21 baseline for smaller-area training screen" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "5d9a47847a3d25bd", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v23-train/scores.tsv", + "research/wake7-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:0223d2e973cd72b9ca4de40402036b530ca0dbc02841115423380aa2b238300b", + "comparison_scope": "ad-placement:local:shrink-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v23-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992089758.1 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T20:40:14Z", + "result_id": "finding_f680cc951186", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "V23 averaged992089758.10: +758611.00 versus matched v21,14/20 wins. Selected before fresh seeds7000..7019 were inspected. All20 strict checks passed.", + "title": "V23 small random area release led training screen" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "3f0e7dcc8dc1f4e4", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v24-train/scores.tsv", + "research/wake7-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8c231509e43e9a809766cd2ab8585969378d681c60cc1769aea8a517cc1788c5", + "comparison_scope": "ad-placement:local:shrink-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v24-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991364503.2 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T20:40:15Z", + "result_id": "finding_16a7e659211d", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "V24 averaged991364503.20: +33356.10 versus matched v21,11/20 wins. All20 strict checks passed. Lower mean than selected v23; no general route refutation.", + "title": "V24 larger random area release gave little training gain" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "26d9cf958a0f83d0", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v25-train/scores.tsv", + "research/wake7-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:13923b8454e2bb94ed32cb7074a3c9c1d4ceca7ea9f710c1710881fc82edffe7", + "comparison_scope": "ad-placement:local:shrink-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v25-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991656786.3 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T20:40:16Z", + "result_id": "finding_f5c19da09934", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "V25 averaged991656786.30: +325639.20 versus matched v21 but only8/20 wins. All20 strict checks passed. It was not selected for the fresh screen.", + "title": "V25 fixed area release had mixed training evidence" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "62c035523f2d3819", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v21-wake7-heldout/scores.tsv", + "research/wake7-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:6b580fbd4974af01ea009b62925a510c6c15dabc74476b9f007ef1bce00dfafd", + "comparison_scope": "ad-placement:local:wake7-heldout:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v21-wake7-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:4da44941af1bc441be2a5b4441e14a20cd2813cdc3a3a0721d41af2503e34b9a", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991200400 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T20:40:18Z", + "result_id": "finding_4c2ea2587e91", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Matched4.70-second baseline averaged991200400 on seeds7000..7019. All20 tester and strict geometry checks passed. Separate comparison scope from earlier cohorts.", + "title": "V21 baseline for fresh smaller-area comparison" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "cd7459992715bff5", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v23-wake7-heldout/scores.tsv", + "research/wake7-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:0223d2e973cd72b9ca4de40402036b530ca0dbc02841115423380aa2b238300b", + "comparison_scope": "ad-placement:local:wake7-heldout:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v23-wake7-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:4da44941af1bc441be2a5b4441e14a20cd2813cdc3a3a0721d41af2503e34b9a", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991269815.9 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T20:40:19Z", + "result_id": "finding_eb2752ab565b", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "V23 averaged991269815.90: +69415.90 versus matched v21,11/20 wins. All20 strict checks passed; max elapsed4.77502s. Gain is small and uncertain; it does not establish native improvement.", + "title": "V23 fresh smaller-area gain was narrow" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "681e8f4b3a614111", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v21-headroom-train/scores.tsv", + "research/wake7-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:6b580fbd4974af01ea009b62925a510c6c15dabc74476b9f007ef1bce00dfafd", + "comparison_scope": "ad-placement:local:headroom-train:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v21-headroom-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991618953.85 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T20:47:47Z", + "result_id": "finding_f657598c676f", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Matched4.70-second baseline averaged991618953.85 on seeds0..19. All20 tester and strict geometry checks passed.", + "title": "V21 baseline for ceiling-based selection training" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "0d5d3ed017cddae6", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v26-train/scores.tsv", + "research/wake7-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:2ceea360ef2c8e2fd81b62791ca9681f6da0e66327d12c186d5f8024a7fcecf5", + "comparison_scope": "ad-placement:local:headroom-train:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v26-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991631729.8 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T20:47:49Z", + "result_id": "finding_d5d7a1967972", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "V26 averaged991631729.80: +12775.95 versus matched v21,11/20 wins. Preprocessing is inside4.70s. All20 strict checks passed; no established quality advantage.", + "title": "V26 ceiling-based selection gave tiny training gain" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "0e24752e1f956f71", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v21-headroom-heldout/scores.tsv", + "research/wake7-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:6b580fbd4974af01ea009b62925a510c6c15dabc74476b9f007ef1bce00dfafd", + "comparison_scope": "ad-placement:local:headroom-heldout:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v21-headroom-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:1783e9d0cbef954941d4edc6957e73a19dd60011b67b2d707f1b91baff862097", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 989691596.9 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T20:47:50Z", + "result_id": "finding_2a987ab9d833", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Matched4.70-second v21 averaged989691596.90 on reserved seeds8000..8019. All20 tester and strict geometry checks passed. This cohort now supports candidate selection.", + "title": "V21 baseline for common fresh candidate selection" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "8455cf6383b7655c", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v23-headroom-heldout/scores.tsv", + "research/wake7-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:0223d2e973cd72b9ca4de40402036b530ca0dbc02841115423380aa2b238300b", + "comparison_scope": "ad-placement:local:headroom-heldout:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v23-headroom-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:1783e9d0cbef954941d4edc6957e73a19dd60011b67b2d707f1b91baff862097", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 989277030.2 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T20:47:51Z", + "result_id": "finding_2ea7fc78b631", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "V23 averaged989277030.20: -414566.70 versus common v21,8/20 wins. All20 strict checks passed. Preserve this scoped regression alongside its earlier narrow positive result.", + "title": "V23 smaller-area proposals regressed on common fresh suite" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "462f498dde045155", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v26-headroom-heldout/scores.tsv", + "research/wake7-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:2ceea360ef2c8e2fd81b62791ca9681f6da0e66327d12c186d5f8024a7fcecf5", + "comparison_scope": "ad-placement:local:headroom-heldout:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v26-headroom-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:1783e9d0cbef954941d4edc6957e73a19dd60011b67b2d707f1b91baff862097", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990006088.8 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T20:47:52Z", + "result_id": "finding_7351d318606e", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "V26 averaged990006088.80: +314491.90 versus v21,10/20 wins. All20 strict checks passed,max4.73011s. Exploratory95% interval[-412291.4,1089552.2] includes zero. Declared rule selects v26 for native evaluation.", + "title": "V26 led common fresh suite with uncertain gain" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "c31c1bf65de8215f", + "event_kind": "node", + "evidence_refs": [ + "research/wake7-summary.json", + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T20:48:31Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 generated tests. V26 led seeds8000-8019 by314492 mean points over v21,10/20 wins; uncertainty is large. V23 regressed there. Native feedback is separate.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "df118607da2aa437", + "event_kind": "node", + "evidence_refs": [ + "research/native/result-v26-poll3.txt", + "research/native/submission-4.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T20:55:07Z", + "result_id": "node_9a8e7c2231a3", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Native fixed50: v3 49443802408, v7 49483706996, v18 49510341240, v26 49554485611; all50/50. V26 is measured incumbent; optimization remains open.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "732816f7acf3ec90", + "event_kind": "node", + "evidence_refs": [ + "research/native/result-v26-poll3.txt", + "research/native/submission-4.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T20:56:45+00:00", + "result_id": "node_9a8e7c2231a3", + "run_id": "explore-result:cc8091ce1a9a2e456fb8f38c2357ca43cba1b10bf17d7df72e456b70a76acf7f", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Native fixed50: v3 49443802408, v7 49483706996, v18 49510341240, v26 49554485611; all50/50. V26 is measured incumbent; optimization remains open.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "8d1626454bbecce4", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-v26-native" + ], + "execution_observation": { + "candidate_sha256": "2ceea360ef2c8e2fd81b62791ca9681f6da0e66327d12c186d5f8024a7fcecf5", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "evidence:ad-v26-native" + ], + "executed_method_id": "sforge.native.fixed50.v1", + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "limit_seconds": null, + "planned_method_id": "sforge.native.fixed50.v1", + "result_candidate_sha256": "2ceea360ef2c8e2fd81b62791ca9681f6da0e66327d12c186d5f8024a7fcecf5", + "result_ref": "evidence:ad-v26-native", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Server completed50/50; per-case execution times, memory and search iterations were not exposed." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:2ceea360ef2c8e2fd81b62791ca9681f6da0e66327d12c186d5f8024a7fcecf5", + "comparison_scope": "ad-placement:native:current-task:fixed50", + "evaluation_ref": "native-protocol:README-sha256:0051d2a6e88e299efc0c0b0f1a7e8a68b3060c855b55ffce8c6cfa601d4b30b2", + "evidence_ref": "evidence:ad-v26-native", + "higher_is_better": true, + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "metric": "total_native_score", + "unit": "points_out_of_50e9", + "value": 49554485611 + } + ], + "node_id": "node_9a8e7c2231a3", + "recorded_at": "2026-10-22T20:56:45+00:00", + "result_id": "finding_cc8091ce1a9a2e456fb8f38c2357ca43cba1b10bf17d7df72e456b70a76acf7f", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50 Applicability: Native fixed50: v3 49443802408, v7 49483706996, v18 49510341240, v26 49554485611; all50/50. V26 is measured incumbent; optimization remains open. Observation: V26/source2ceea360, submission6d242d9e37ce, passed50/50 and scored49554485611: +44144371 versus native v18. Retain the exact accepted snapshot in solution.cpp. Interpretation: V26 is the highest-scoring tested native candidate. It combines current-area proposals and ceiling-based selection relative to v18; native gain does not isolate their contributions. V23 regressed on the common fresh suite, while the additional local v26 gain over v21 was uncertain. Keep all snapshots and original constraints. Next test move allocation from separate diagnostics; no global optimum or Goal completion is claimed.", + "tags": [ + "writeback-result" + ], + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "b5b1f7e9ee4f3556", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v26-allocation-train/scores.tsv", + "research/wake8-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:2ceea360ef2c8e2fd81b62791ca9681f6da0e66327d12c186d5f8024a7fcecf5", + "comparison_scope": "ad-placement:local:allocation-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v26-allocation-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991123759.75 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T21:06:10Z", + "result_id": "finding_56a8d232eefb", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Matched4.70-second baseline averaged991123759.75 on seeds0..19. All20 tester and strict geometry checks passed. Compare within this screen because earlier timed baselines vary.", + "title": "V26 baseline for move-allocation training screen" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "d262e0b4f250f5de", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v27-train/scores.tsv", + "research/wake8-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:e10f96fe1ba0ff9aa27ee8d6f16a85ad3c6c164b366446b9ea3e687395cee492", + "comparison_scope": "ad-placement:local:allocation-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v27-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991574565.55 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T21:06:20Z", + "result_id": "finding_daecbe0901b4", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "V27 averaged991574565.55: +450805.80 versus matched v26,13/20 wins. Growth30%,local aspect37%; chain frequency and fill semantics preserved. All20 strict checks passed; selected before either fresh suite was scored.", + "title": "V27 moderate aspect allocation led training screen" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "7a0515b5980fd725", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v28-train/scores.tsv", + "research/wake8-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:ed030ca85e38f3b61654e6a7918658422c9186a1578bd114e4087f6ab1a88623", + "comparison_scope": "ad-placement:local:allocation-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v28-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991390784.75 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T21:06:24Z", + "result_id": "finding_3a5e8b375a0b", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "V28 averaged991390784.75: +267025.00 versus matched v26,10/20 wins. Growth20%,local aspect47%; all20 strict checks passed. Lower mean than v27 does not refute this broader route.", + "title": "V28 larger aspect allocation had mixed training evidence" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "eb417269426ad8dc", + "event_kind": "finding", + "evidence_refs": [ + "research/allocation-v27-heldout40-comparison.json", + "research/wake8-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:2ceea360ef2c8e2fd81b62791ca9681f6da0e66327d12c186d5f8024a7fcecf5", + "comparison_scope": "ad-placement:local:allocation-heldout40:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/allocation-v27-heldout40-comparison.json", + "higher_is_better": true, + "input_revision": "sha256:fa3f5a9a2b2a7d07a6c9a3553860e18afb40f9171917f2e472cec16fcbba2179", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991334106.275 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T21:15:36Z", + "result_id": "finding_3e85cfd6d138", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Matched4.70-second v26 averaged991334106.275 across both reserved20-case cohorts. All40 tester and independent strict geometry checks passed. Both cohorts remain included.", + "title": "V26 baseline for40-case fresh allocation comparison" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "d160c9695cf6c6a6", + "event_kind": "node", + "evidence_refs": [ + "research/wake8-summary.json", + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T21:15:37Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 generated tests. V27 gained143638 mean points over v26 on40 reserved cases,22/40 wins; uncertainty is large. V27 is staged; v26 remains native incumbent.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "fa3225f5089dcf19", + "event_kind": "node", + "evidence_refs": [ + "research/wake8-summary.json", + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T21:18:08+00:00", + "result_id": "node_e388daf41666", + "run_id": "explore-result:937fede4f1188f5004b4797c3aab4f02481063178441b17940b02cdc0c04b0cf", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 generated tests. V27 gained143638 mean points over v26 on40 reserved cases,22/40 wins; uncertainty is large. V27 is staged; v26 remains native incumbent.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "ea7e19f9d9a5046e", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-v27-allocation40" + ], + "execution_observation": { + "candidate_sha256": "e10f96fe1ba0ff9aa27ee8d6f16a85ad3c6c164b366446b9ea3e687395cee492", + "change_reason": null, + "elapsed_seconds": 189.53577, + "evidence_refs": [ + "evidence:ad-v27-allocation40" + ], + "executed_method_id": "ad-placement.paired.40x4p70s.v1", + "exit_code": 0, + "input_revision": "sha256:fa3f5a9a2b2a7d07a6c9a3553860e18afb40f9171917f2e472cec16fcbba2179", + "limit_seconds": 200, + "planned_method_id": "ad-placement.paired.40x4p70s.v1", + "result_candidate_sha256": "e10f96fe1ba0ff9aa27ee8d6f16a85ad3c6c164b366446b9ea3e687395cee492", + "result_ref": "evidence:ad-v27-allocation40", + "schema_version": "explore_execution_observation_v0", + "search_iterations": 238992936, + "stop_reason": "40 serial searches completed; each used4.70-second search budget within5-second case limit." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:e10f96fe1ba0ff9aa27ee8d6f16a85ad3c6c164b366446b9ea3e687395cee492", + "comparison_scope": "ad-placement:local:allocation-heldout40:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "evidence:ad-v27-allocation40", + "higher_is_better": true, + "input_revision": "sha256:fa3f5a9a2b2a7d07a6c9a3553860e18afb40f9171917f2e472cec16fcbba2179", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991477744.025 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T21:18:08+00:00", + "result_id": "finding_937fede4f1188f5004b4797c3aab4f02481063178441b17940b02cdc0c04b0cf", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:fa3f5a9a2b2a7d07a6c9a3553860e18afb40f9171917f2e472cec16fcbba2179 Applicability: 4.70-second O2 generated tests. V27 gained143638 mean points over v26 on40 reserved cases,22/40 wins; uncertainty is large. V27 is staged; v26 remains native incumbent. Observation: V27 averaged991477744.025 on40 fresh cases: +143637.75 versus v26,22/40 wins. All40 strict checks passed; candidate maximum4.78249s. Sourcee10f96fe is unsubmitted. Interpretation: V27 was selected on training before either fresh suite was scored. Primary gain5486 and extra gain281789 points remain combined; exploratory95% interval[-303111.6,591378.125] includes zero. Stage v27 while retaining native v26. V28 had a lower training mean; this does not refute allocation tuning. Use eligible source-matched native evaluation and preserve original constraints. Execution fields are aggregate40-case totals, not per-case times.", + "tags": [ + "writeback-result" + ], + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "8f2a138f94d4b72a", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v27-contact-train/scores.tsv", + "research/wake9-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:e10f96fe1ba0ff9aa27ee8d6f16a85ad3c6c164b366446b9ea3e687395cee492", + "comparison_scope": "ad-placement:local:contact-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v27-contact-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990887676.1 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T21:44:37Z", + "result_id": "finding_7b6d4979fc34", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v27 source e10f96fe averaged 990887676.1 points on20 original-score/O2/4.70-second cases. Scope: contact training. All tester and independent geometry checks passed. This local result does not establish native gain.", + "title": "Wake9 v27 contact training mean 990887676.1" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "08df3cd5a9db57b5", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v29-train/scores.tsv", + "research/wake9-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:d8cb40343d22954b54f9b76bc649b13b6c607630c2e9310df13e517086d5422e", + "comparison_scope": "ad-placement:local:contact-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v29-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991112119.85 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T21:44:39Z", + "result_id": "finding_315902cd249c", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v29 source d8cb4034 averaged 991112119.85 points on20 original-score/O2/4.70-second cases. Scope: contact training. All tester and independent geometry checks passed. This local result does not establish native gain.", + "title": "Wake9 v29 contact training mean 991112119.85" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "4176a0c162c2e51d", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v30-train/scores.tsv", + "research/wake9-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:0e2c060097341ab803ca9287a71614b9f68a1b6e8b12d54ac47813dfa772b562", + "comparison_scope": "ad-placement:local:contact-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v30-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991783632.2 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T21:44:41Z", + "result_id": "finding_c052ad2eec09", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v30 source 0e2c0600 averaged 991783632.2 points on20 original-score/O2/4.70-second cases. Scope: contact training. All tester and independent geometry checks passed. This local result does not establish native gain.", + "title": "Wake9 v30 contact training mean 991783632.2" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "352fcdaf449c7fbe", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v27-contact-heldout/scores.tsv", + "research/wake9-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:e10f96fe1ba0ff9aa27ee8d6f16a85ad3c6c164b366446b9ea3e687395cee492", + "comparison_scope": "ad-placement:local:contact-heldout:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v27-contact-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:49654c92787dc3705b0b8b7ba0dba1a4d333e82000739422b8f992e37b9aa3ed", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990128870.4 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T21:44:43Z", + "result_id": "finding_d3f3df79905b", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v27 source e10f96fe averaged 990128870.4 points on20 original-score/O2/4.70-second cases. Scope: contact primary fresh. All tester and independent geometry checks passed. This local result does not establish native gain.", + "title": "Wake9 v27 contact primary fresh mean 990128870.4" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "a4b5a75294a67083", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v30-contact-heldout/scores.tsv", + "research/wake9-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:0e2c060097341ab803ca9287a71614b9f68a1b6e8b12d54ac47813dfa772b562", + "comparison_scope": "ad-placement:local:contact-heldout:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v30-contact-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:49654c92787dc3705b0b8b7ba0dba1a4d333e82000739422b8f992e37b9aa3ed", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990310535.15 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T21:44:45Z", + "result_id": "finding_cd7a0e33b821", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v30 source 0e2c0600 averaged 990310535.15 points on20 original-score/O2/4.70-second cases. Scope: contact primary fresh. All tester and independent geometry checks passed. This local result does not establish native gain.", + "title": "Wake9 v30 contact primary fresh mean 990310535.15" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "93c47390c04b4827", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v27-warm-train/scores.tsv", + "research/wake9-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:e10f96fe1ba0ff9aa27ee8d6f16a85ad3c6c164b366446b9ea3e687395cee492", + "comparison_scope": "ad-placement:local:warm-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v27-warm-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991867314.15 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T21:44:47Z", + "result_id": "finding_f28318b2ce95", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v27 source e10f96fe averaged 991867314.15 points on20 original-score/O2/4.70-second cases. Scope: warm training. All tester and independent geometry checks passed. This local result does not establish native gain.", + "title": "Wake9 v27 warm training mean 991867314.15" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "2973dbd891c7180a", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v31-train/scores.tsv", + "research/wake9-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:e1b576e50780385f93dca1663d757039190dff3ebbee315125234155ccd5930e", + "comparison_scope": "ad-placement:local:warm-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v31-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991093474.8 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T21:44:49Z", + "result_id": "finding_b4be73406c3b", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v31 source e1b576e5 averaged 991093474.8 points on20 original-score/O2/4.70-second cases. Scope: warm training. All tester and independent geometry checks passed. This local result does not establish native gain.", + "title": "Wake9 v31 warm training mean 991093474.8" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "c81024e480ead688", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v32-train/scores.tsv", + "research/wake9-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:59ed006a1bbbb0c3f3047f25f9cb0a5b0d29a8808f003f690c4ac724e69303a6", + "comparison_scope": "ad-placement:local:warm-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v32-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991796393.55 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T21:44:50Z", + "result_id": "finding_4730c5b18b08", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v32 source 59ed006a averaged 991796393.55 points on20 original-score/O2/4.70-second cases. Scope: warm training. All tester and independent geometry checks passed. This local result does not establish native gain.", + "title": "Wake9 v32 warm training mean 991796393.55" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "333ed715d2286645", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v27-warm-heldout/scores.tsv", + "research/wake9-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:e10f96fe1ba0ff9aa27ee8d6f16a85ad3c6c164b366446b9ea3e687395cee492", + "comparison_scope": "ad-placement:local:warm-heldout:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v27-warm-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:ed57dd3dcd24587dcf40ea7d96c96a13b31ae22e084b3ef72c14648e1ca304e2", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992177093.65 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T21:44:52Z", + "result_id": "finding_929f2352dae2", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v27 source e10f96fe averaged 992177093.65 points on20 original-score/O2/4.70-second cases. Scope: warm primary fresh. All tester and independent geometry checks passed. This local result does not establish native gain.", + "title": "Wake9 v27 warm primary fresh mean 992177093.65" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "66494e6398498f4c", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v32-warm-heldout/scores.tsv", + "research/wake9-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:59ed006a1bbbb0c3f3047f25f9cb0a5b0d29a8808f003f690c4ac724e69303a6", + "comparison_scope": "ad-placement:local:warm-heldout:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v32-warm-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:ed57dd3dcd24587dcf40ea7d96c96a13b31ae22e084b3ef72c14648e1ca304e2", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991230484.9 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T21:44:54Z", + "result_id": "finding_a6a435ff29c8", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v32 source 59ed006a averaged 991230484.9 points on20 original-score/O2/4.70-second cases. Scope: warm primary fresh. All tester and independent geometry checks passed. This local result does not establish native gain.", + "title": "Wake9 v32 warm primary fresh mean 991230484.9" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "912aec213cf60255", + "event_kind": "finding", + "evidence_refs": [ + "research/native/result-v27-poll2.txt", + "research/native/v27-submission.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:e10f96fe1ba0ff9aa27ee8d6f16a85ad3c6c164b366446b9ea3e687395cee492", + "comparison_scope": "ad-placement:native:current-task:fixed50", + "evaluation_ref": "native-protocol:README-sha256:0051d2a6e88e299efc0c0b0f1a7e8a68b3060c855b55ffce8c6cfa601d4b30b2", + "evidence_ref": "research/native/result-v27-poll2.txt", + "higher_is_better": true, + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "metric": "total_native_score", + "unit": "points_out_of_50e9", + "value": 49549612741 + } + ], + "node_id": "node_9a8e7c2231a3", + "recorded_at": "2026-10-22T21:53:04Z", + "result_id": "finding_cfc74dff8614", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Accepted source e10f96fe in submission 9ff36082941c scored49549612741,50/50:4872870 below nativev26. Exactv26/source2ceea360 restored. Local gains did not establish native improvement. Original constraints and optimization remain active.", + "title": "Native v27 trial scores 49549612741 with 50/50 passes; retain v26" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "ea8e690ae61066f3", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T21:53:23Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 tests. Contact v30 gained181665 on20 fresh cases (9wins); partial-restart v32 lost946609 (7wins). Neither displaced v27. Native incumbent v26 is retained.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "ed7ba3e0be76dc47", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T21:53:24Z", + "result_id": "node_9a8e7c2231a3", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Native fixed50: v26 49554485611 (50/50), v27 49549612741 (50/50). v26 is the measured incumbent; source-matched feedback and controlled optimization continue.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "1c112354dbf09b89", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T21:55:15+00:00", + "result_id": "node_9a8e7c2231a3", + "run_id": "explore-result:2c2ca4c86aea1c895d44ac8603a74a20dc849e6d9fe6b506a0b3c2c9e133a8b6", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Native fixed50: v26 49554485611 (50/50), v27 49549612741 (50/50). v26 is the measured incumbent; source-matched feedback and controlled optimization continue.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "42f1d98d9ca16180", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-v27-native" + ], + "execution_observation": { + "candidate_sha256": "e10f96fe1ba0ff9aa27ee8d6f16a85ad3c6c164b366446b9ea3e687395cee492", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "evidence:ad-v27-native" + ], + "executed_method_id": "ad-placement.native.fixed50.v1", + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "limit_seconds": null, + "planned_method_id": "ad-placement.native.fixed50.v1", + "result_candidate_sha256": "e10f96fe1ba0ff9aa27ee8d6f16a85ad3c6c164b366446b9ea3e687395cee492", + "result_ref": "evidence:ad-v27-native", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Accepted native evaluation completed; resource details are not exposed." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:e10f96fe1ba0ff9aa27ee8d6f16a85ad3c6c164b366446b9ea3e687395cee492", + "comparison_scope": "ad-placement:native:current-task:fixed50", + "evaluation_ref": "native-protocol:README-sha256:0051d2a6e88e299efc0c0b0f1a7e8a68b3060c855b55ffce8c6cfa601d4b30b2", + "evidence_ref": "evidence:ad-v27-native", + "higher_is_better": true, + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "metric": "total_native_score", + "unit": "points_out_of_50e9", + "value": 49549612741 + } + ], + "node_id": "node_9a8e7c2231a3", + "recorded_at": "2026-10-22T21:55:15+00:00", + "result_id": "finding_2c2ca4c86aea1c895d44ac8603a74a20dc849e6d9fe6b506a0b3c2c9e133a8b6", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50 Applicability: Native fixed50: v26 49554485611 (50/50), v27 49549612741 (50/50). v26 is the measured incumbent; source-matched feedback and controlled optimization continue. Observation: Frozen v27/sourcee10f96fe scored49549612741 with50/50 passes in submission9ff36082941c. Delta versus nativev26:-4872870. Retained v26. Interpretation: The accepted archive and result ID match the frozen source. V27 did not beat the valid native incumbent; exactv26 was restored. Local gains did not certify native improvement. All public per-test statuses are preserved, but per-case scores, execution times, memory and iteration counts were not exposed. Null resource fields mean unmeasured. Original constraints and optimization remain active; next is the separate cooling probe.", + "tags": [ + "writeback-result" + ], + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "e3a7c1cb4fb78776", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v26-cooling-train/scores.tsv", + "research/wake10-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:2ceea360ef2c8e2fd81b62791ca9681f6da0e66327d12c186d5f8024a7fcecf5", + "comparison_scope": "ad-placement:local:cooling-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v26-cooling-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991205498.9 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:14:15Z", + "result_id": "finding_e92cf27f01b2", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v26 source2ceea360 averaged991205498.9 on20 original-score/O2/4.70-second generated cases. Scope:train. All tester and independent geometry checks passed. Native gain remains unmeasured.", + "title": "Cooling v26 train mean 991205498.9" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "3d96be8e396738b9", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v33-cooling-train/scores.tsv", + "research/wake10-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:2375b224819546a15fb839632f40feaaf8bfc4f853f7a145be74aabf374a01b9", + "comparison_scope": "ad-placement:local:cooling-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v33-cooling-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991381920.25 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:14:18Z", + "result_id": "finding_7458b3136205", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v33 source2375b224 averaged991381920.25 on20 original-score/O2/4.70-second generated cases. Scope:train. All tester and independent geometry checks passed. Native gain remains unmeasured.", + "title": "Cooling v33 train mean 991381920.25" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "fba53c0c07f0d7a9", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v34-cooling-train/scores.tsv", + "research/wake10-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:59221336f2421862b527e336b3922b7910d5ecb335a516f92e6e9df493a13acd", + "comparison_scope": "ad-placement:local:cooling-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v34-cooling-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992513331.55 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:14:20Z", + "result_id": "finding_71750ad384ec", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v34 source59221336 averaged992513331.55 on20 original-score/O2/4.70-second generated cases. Scope:train. All tester and independent geometry checks passed. Native gain remains unmeasured.", + "title": "Cooling v34 train mean 992513331.55" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "9c9e494f2b00d208", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v26-cooling-heldout/scores.tsv", + "research/wake10-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:2ceea360ef2c8e2fd81b62791ca9681f6da0e66327d12c186d5f8024a7fcecf5", + "comparison_scope": "ad-placement:local:cooling-heldout:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v26-cooling-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:696abda6bd1d36c95da155090a5dfaa9146a519ad73d5b209db539fcf3003ca8", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992037175.5 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:14:21Z", + "result_id": "finding_62da32038f1f", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v26 source2ceea360 averaged992037175.5 on20 original-score/O2/4.70-second generated cases. Scope:heldout. All tester and independent geometry checks passed. Native gain remains unmeasured.", + "title": "Cooling v26 heldout mean 992037175.5" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "3abe700a1a17bcaa", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v34-cooling-heldout/scores.tsv", + "research/wake10-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:59221336f2421862b527e336b3922b7910d5ecb335a516f92e6e9df493a13acd", + "comparison_scope": "ad-placement:local:cooling-heldout:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v34-cooling-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:696abda6bd1d36c95da155090a5dfaa9146a519ad73d5b209db539fcf3003ca8", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992185371.2 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:14:22Z", + "result_id": "finding_f9363563a642", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v34 source59221336 averaged992185371.2 on20 original-score/O2/4.70-second generated cases. Scope:heldout. All tester and independent geometry checks passed. Native gain remains unmeasured.", + "title": "Cooling v34 heldout mean 992185371.2" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "482e31b33bc80dea", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v26-cooling-extra/scores.tsv", + "research/wake10-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:2ceea360ef2c8e2fd81b62791ca9681f6da0e66327d12c186d5f8024a7fcecf5", + "comparison_scope": "ad-placement:local:cooling-extra:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v26-cooling-extra/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:e5c13d4548cdef35bd68596c1a281c56bf346de1e66a957b518314a29cd3b01f", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991259693.25 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:14:23Z", + "result_id": "finding_8dbfd7f00d14", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v26 source2ceea360 averaged991259693.25 on20 original-score/O2/4.70-second generated cases. Scope:extra. All tester and independent geometry checks passed. Native gain remains unmeasured.", + "title": "Cooling v26 extra mean 991259693.25" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "97f6a7f681108753", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v34-cooling-extra/scores.tsv", + "research/wake10-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:59221336f2421862b527e336b3922b7910d5ecb335a516f92e6e9df493a13acd", + "comparison_scope": "ad-placement:local:cooling-extra:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v34-cooling-extra/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:e5c13d4548cdef35bd68596c1a281c56bf346de1e66a957b518314a29cd3b01f", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992018288.35 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:14:24Z", + "result_id": "finding_8d91bb9a49ed", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v34 source59221336 averaged992018288.35 on20 original-score/O2/4.70-second generated cases. Scope:extra. All tester and independent geometry checks passed. Native gain remains unmeasured.", + "title": "Cooling v34 extra mean 992018288.35" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "8088c3983054013b", + "event_kind": "finding", + "evidence_refs": [ + "research/cooling-fresh40-summary.json", + "research/wake10-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:2ceea360ef2c8e2fd81b62791ca9681f6da0e66327d12c186d5f8024a7fcecf5", + "comparison_scope": "ad-placement:local:cooling-fresh40:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/cooling-fresh40-summary.json", + "higher_is_better": true, + "input_revision": "sha256:bce985c8588d7453b83048c2e1c8913f4f3bf51bb0521ce57462e94ecc477d35", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991648434.375 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:14:25Z", + "result_id": "finding_e3752015b3f7", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v26 source2ceea360 averaged991648434.375 on40 original-score/O2/4.70-second generated cases. Scope:fresh40. All tester and independent geometry checks passed. Native gain remains unmeasured.", + "title": "Cooling v26 fresh40 mean 991648434.375" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "fd098e432c9621cc", + "event_kind": "finding", + "evidence_refs": [ + "research/cooling-fresh40-summary.json", + "research/wake10-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:59221336f2421862b527e336b3922b7910d5ecb335a516f92e6e9df493a13acd", + "comparison_scope": "ad-placement:local:cooling-fresh40:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/cooling-fresh40-summary.json", + "higher_is_better": true, + "input_revision": "sha256:bce985c8588d7453b83048c2e1c8913f4f3bf51bb0521ce57462e94ecc477d35", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992101829.775 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:14:26Z", + "result_id": "finding_489cdd1cba17", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v34 source59221336 averaged992101829.775 on40 original-score/O2/4.70-second generated cases. Scope:fresh40. All tester and independent geometry checks passed. Native gain remains unmeasured.", + "title": "Cooling v34 fresh40 mean 992101829.775" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "b155e361d3b7c9f1", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T22:15:26Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 tests. V34 gained453395 mean points on40 fresh cases (24wins); uncertainty retained. Staged for native; nativev26 remains incumbent.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "324f437fa8598322", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T22:17:47+00:00", + "result_id": "node_e388daf41666", + "run_id": "explore-result:f95d6e3d34256e6a2a915686c22ee5052cc3924b5df0339be76fa1107f4f1dda", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 tests. V34 gained453395 mean points on40 fresh cases (24wins); uncertainty retained. Staged for native; nativev26 remains incumbent.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "bb187f961de63c4b", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-cooling-fresh40" + ], + "execution_observation": { + "candidate_sha256": "59221336f2421862b527e336b3922b7910d5ecb335a516f92e6e9df493a13acd", + "change_reason": null, + "elapsed_seconds": 188.94440000000003, + "evidence_refs": [ + "evidence:ad-cooling-fresh40" + ], + "executed_method_id": "ad-placement.paired.40x4p70s.v1", + "exit_code": 0, + "input_revision": "sha256:bce985c8588d7453b83048c2e1c8913f4f3bf51bb0521ce57462e94ecc477d35", + "limit_seconds": 200, + "planned_method_id": "ad-placement.paired.40x4p70s.v1", + "result_candidate_sha256": "59221336f2421862b527e336b3922b7910d5ecb335a516f92e6e9df493a13acd", + "result_ref": "evidence:ad-cooling-fresh40", + "schema_version": "explore_execution_observation_v0", + "search_iterations": 271292712, + "stop_reason": "All40 serial searches completed within5seconds each; both predeclared cohorts retained." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:59221336f2421862b527e336b3922b7910d5ecb335a516f92e6e9df493a13acd", + "comparison_scope": "ad-placement:local:cooling-fresh40:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "evidence:ad-cooling-fresh40", + "higher_is_better": true, + "input_revision": "sha256:bce985c8588d7453b83048c2e1c8913f4f3bf51bb0521ce57462e94ecc477d35", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992101829.775 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:17:47+00:00", + "result_id": "finding_f95d6e3d34256e6a2a915686c22ee5052cc3924b5df0339be76fa1107f4f1dda", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:bce985c8588d7453b83048c2e1c8913f4f3bf51bb0521ce57462e94ecc477d35 Applicability: 4.70-second O2 tests. V34 gained453395 mean points on40 fresh cases (24wins); uncertainty retained. Staged for native; nativev26 remains incumbent. Observation: V34 averaged992101829.775 on40 fresh cases: delta453395.4 versus v26,24/40wins. All40 checks passed; maximum4.74753s. Native result is unmeasured. Interpretation: The training-selected source was frozen before either predeclared cohort. Both cohorts remain combined. Exploratory95% paired-case interval[9391.225,921642.2] does not account for scheduling noise, selection multiplicity or native shift. Stage the positive-mean candidate while retaining exact nativev26. Diagnostics confirm reached temperature/acceptance bins, not quality rankings. Execution fields aggregate40 full-budget cases, not one case. Next probe is separate and unimplemented.", + "tags": [ + "writeback-result" + ], + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "f4fbfa4da7ae5e64", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v26-uppertemp-train/scores.tsv", + "research/wake11-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:2ceea360ef2c8e2fd81b62791ca9681f6da0e66327d12c186d5f8024a7fcecf5", + "comparison_scope": "ad-placement:local:uppertemp-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v26-uppertemp-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991554009.8 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:45:04Z", + "result_id": "finding_10c30602025c", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v26 source 2ceea360 averaged 991554009.8 on 20 original-score/O2/4.70-second generated cases. Scope: train. All tester and independent geometry checks passed. Native gain remains unmeasured at this local finding.", + "title": "Temperature v26 train mean 991554009.8" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "f51b14338f697c90", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v34-uppertemp-train/scores.tsv", + "research/wake11-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:59221336f2421862b527e336b3922b7910d5ecb335a516f92e6e9df493a13acd", + "comparison_scope": "ad-placement:local:uppertemp-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v34-uppertemp-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992599352 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:45:07Z", + "result_id": "finding_e618d80c851c", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v34 source 59221336 averaged 992599352 on 20 original-score/O2/4.70-second generated cases. Scope: train. All tester and independent geometry checks passed. Native gain remains unmeasured at this local finding.", + "title": "Temperature v34 train mean 992599352" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "e7a1918599e28917", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v35-uppertemp-train/scores.tsv", + "research/wake11-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:uppertemp-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v35-uppertemp-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992607171.3 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:45:08Z", + "result_id": "finding_c6f21001ae09", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 source 8af4a746 averaged 992607171.3 on 20 original-score/O2/4.70-second generated cases. Scope: train. All tester and independent geometry checks passed. Native gain remains unmeasured at this local finding.", + "title": "Temperature v35 train mean 992607171.3" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "7340ee69ed758e7d", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v36-uppertemp-train/scores.tsv", + "research/wake11-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:40f4fe30ac39c1db125199cda06f796123d0d5c0e11e8d2774fecb77498edc30", + "comparison_scope": "ad-placement:local:uppertemp-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v36-uppertemp-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992111453.5 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:45:10Z", + "result_id": "finding_9ceabd7346d5", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v36 source 40f4fe30 averaged 992111453.5 on 20 original-score/O2/4.70-second generated cases. Scope: train. All tester and independent geometry checks passed. Native gain remains unmeasured at this local finding.", + "title": "Temperature v36 train mean 992111453.5" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "69a18f8847cdd44b", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v26-uppertemp-heldout/scores.tsv", + "research/wake11-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:2ceea360ef2c8e2fd81b62791ca9681f6da0e66327d12c186d5f8024a7fcecf5", + "comparison_scope": "ad-placement:local:uppertemp-heldout:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v26-uppertemp-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:bc97f1124c8783e8d9c212b40d34f7fd35ac5483ba3390d6776192caeabeeec9", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992146859.1 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:45:11Z", + "result_id": "finding_a75f7109a34a", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v26 source 2ceea360 averaged 992146859.1 on 20 original-score/O2/4.70-second generated cases. Scope: heldout. All tester and independent geometry checks passed. Native gain remains unmeasured at this local finding.", + "title": "Temperature v26 heldout mean 992146859.1" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "e99f1dfec81e5dde", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v34-uppertemp-heldout/scores.tsv", + "research/wake11-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:59221336f2421862b527e336b3922b7910d5ecb335a516f92e6e9df493a13acd", + "comparison_scope": "ad-placement:local:uppertemp-heldout:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v34-uppertemp-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:bc97f1124c8783e8d9c212b40d34f7fd35ac5483ba3390d6776192caeabeeec9", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992773602.85 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:45:12Z", + "result_id": "finding_2d3483541866", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v34 source 59221336 averaged 992773602.85 on 20 original-score/O2/4.70-second generated cases. Scope: heldout. All tester and independent geometry checks passed. Native gain remains unmeasured at this local finding.", + "title": "Temperature v34 heldout mean 992773602.85" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "af209ca9131d87c6", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v35-uppertemp-heldout/scores.tsv", + "research/wake11-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:uppertemp-heldout:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v35-uppertemp-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:bc97f1124c8783e8d9c212b40d34f7fd35ac5483ba3390d6776192caeabeeec9", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992518443.25 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:45:14Z", + "result_id": "finding_0877cea57060", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 source 8af4a746 averaged 992518443.25 on 20 original-score/O2/4.70-second generated cases. Scope: heldout. All tester and independent geometry checks passed. Native gain remains unmeasured at this local finding.", + "title": "Temperature v35 heldout mean 992518443.25" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "1563d443ef35c238", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v36-uppertemp-heldout/scores.tsv", + "research/wake11-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:40f4fe30ac39c1db125199cda06f796123d0d5c0e11e8d2774fecb77498edc30", + "comparison_scope": "ad-placement:local:uppertemp-heldout:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v36-uppertemp-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:bc97f1124c8783e8d9c212b40d34f7fd35ac5483ba3390d6776192caeabeeec9", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992681050.25 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:45:15Z", + "result_id": "finding_79943bea5e4f", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v36 source 40f4fe30 averaged 992681050.25 on 20 original-score/O2/4.70-second generated cases. Scope: heldout. All tester and independent geometry checks passed. Native gain remains unmeasured at this local finding.", + "title": "Temperature v36 heldout mean 992681050.25" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "17abc0e2b0357fc1", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v26-uppertemp-extra/scores.tsv", + "research/wake11-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:2ceea360ef2c8e2fd81b62791ca9681f6da0e66327d12c186d5f8024a7fcecf5", + "comparison_scope": "ad-placement:local:uppertemp-extra:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v26-uppertemp-extra/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:5ad00d227827f721668f0ff88e8eb975d50e42e760983ed079ccb7e06aacd25c", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 989031830.4 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:45:17Z", + "result_id": "finding_f61822c9c663", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v26 source 2ceea360 averaged 989031830.4 on 20 original-score/O2/4.70-second generated cases. Scope: extra. All tester and independent geometry checks passed. Native gain remains unmeasured at this local finding.", + "title": "Temperature v26 extra mean 989031830.4" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "7b8d2f06c43a13be", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v34-uppertemp-extra/scores.tsv", + "research/wake11-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:59221336f2421862b527e336b3922b7910d5ecb335a516f92e6e9df493a13acd", + "comparison_scope": "ad-placement:local:uppertemp-extra:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v34-uppertemp-extra/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:5ad00d227827f721668f0ff88e8eb975d50e42e760983ed079ccb7e06aacd25c", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 989707714.7 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:45:18Z", + "result_id": "finding_7422ad4a83ec", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v34 source 59221336 averaged 989707714.7 on 20 original-score/O2/4.70-second generated cases. Scope: extra. All tester and independent geometry checks passed. Native gain remains unmeasured at this local finding.", + "title": "Temperature v34 extra mean 989707714.7" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "dbd399a6a46bcb42", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v35-uppertemp-extra/scores.tsv", + "research/wake11-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:uppertemp-extra:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v35-uppertemp-extra/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:5ad00d227827f721668f0ff88e8eb975d50e42e760983ed079ccb7e06aacd25c", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990907262.25 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:45:20Z", + "result_id": "finding_b725a2131dfa", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 source 8af4a746 averaged 990907262.25 on 20 original-score/O2/4.70-second generated cases. Scope: extra. All tester and independent geometry checks passed. Native gain remains unmeasured at this local finding.", + "title": "Temperature v35 extra mean 990907262.25" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "378e537c5585b52b", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v36-uppertemp-extra/scores.tsv", + "research/wake11-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:40f4fe30ac39c1db125199cda06f796123d0d5c0e11e8d2774fecb77498edc30", + "comparison_scope": "ad-placement:local:uppertemp-extra:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v36-uppertemp-extra/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:5ad00d227827f721668f0ff88e8eb975d50e42e760983ed079ccb7e06aacd25c", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990694116.95 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:45:21Z", + "result_id": "finding_dfb3a1f71034", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v36 source 40f4fe30 averaged 990694116.95 on 20 original-score/O2/4.70-second generated cases. Scope: extra. All tester and independent geometry checks passed. Native gain remains unmeasured at this local finding.", + "title": "Temperature v36 extra mean 990694116.95" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "52d4d12450449df7", + "event_kind": "finding", + "evidence_refs": [ + "research/uppertemp-fresh40-summary.json", + "research/wake11-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:2ceea360ef2c8e2fd81b62791ca9681f6da0e66327d12c186d5f8024a7fcecf5", + "comparison_scope": "ad-placement:local:uppertemp-fresh40:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/uppertemp-fresh40-summary.json", + "higher_is_better": true, + "input_revision": "sha256:3f1e62ac9fe4b2c58600d842d4988637bc332bd428dfb53bf034161a95af4973", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990589344.75 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:45:23Z", + "result_id": "finding_104059a2ffdd", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v26 source 2ceea360 averaged 990589344.75 on 40 original-score/O2/4.70-second generated cases. Scope: fresh40. All tester and independent geometry checks passed. Native gain remains unmeasured at this local finding.", + "title": "Temperature v26 fresh40 mean 990589344.75" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "78399c5a144d72eb", + "event_kind": "finding", + "evidence_refs": [ + "research/uppertemp-fresh40-summary.json", + "research/wake11-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:59221336f2421862b527e336b3922b7910d5ecb335a516f92e6e9df493a13acd", + "comparison_scope": "ad-placement:local:uppertemp-fresh40:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/uppertemp-fresh40-summary.json", + "higher_is_better": true, + "input_revision": "sha256:3f1e62ac9fe4b2c58600d842d4988637bc332bd428dfb53bf034161a95af4973", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991240658.775 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:45:24Z", + "result_id": "finding_249e419a606e", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v34 source 59221336 averaged 991240658.775 on 40 original-score/O2/4.70-second generated cases. Scope: fresh40. All tester and independent geometry checks passed. Native gain remains unmeasured at this local finding.", + "title": "Temperature v34 fresh40 mean 991240658.775" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "e47f9470bc9cbc5b", + "event_kind": "finding", + "evidence_refs": [ + "research/uppertemp-fresh40-summary.json", + "research/wake11-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:uppertemp-fresh40:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/uppertemp-fresh40-summary.json", + "higher_is_better": true, + "input_revision": "sha256:3f1e62ac9fe4b2c58600d842d4988637bc332bd428dfb53bf034161a95af4973", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991712852.75 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:45:25Z", + "result_id": "finding_7d61cc5ce5dd", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 source 8af4a746 averaged 991712852.75 on 40 original-score/O2/4.70-second generated cases. Scope: fresh40. All tester and independent geometry checks passed. Native gain remains unmeasured at this local finding.", + "title": "Temperature v35 fresh40 mean 991712852.75" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "28cbee74a79dc199", + "event_kind": "finding", + "evidence_refs": [ + "research/uppertemp-fresh40-summary.json", + "research/wake11-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:40f4fe30ac39c1db125199cda06f796123d0d5c0e11e8d2774fecb77498edc30", + "comparison_scope": "ad-placement:local:uppertemp-fresh40:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/uppertemp-fresh40-summary.json", + "higher_is_better": true, + "input_revision": "sha256:3f1e62ac9fe4b2c58600d842d4988637bc332bd428dfb53bf034161a95af4973", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991687583.6 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T22:45:27Z", + "result_id": "finding_92c5bcf2c293", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v36 source 40f4fe30 averaged 991687583.6 on 40 original-score/O2/4.70-second generated cases. Scope: fresh40. All tester and independent geometry checks passed. Native gain remains unmeasured at this local finding.", + "title": "Temperature v36 fresh40 mean 991687583.6" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "82606fc066f0008b", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T22:53:12Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 tests. v35 selected on40 common fresh cases, delta1123508 vs v26. Selection uncertainty retained. Native incumbent v35.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "3107cf55b9850092", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T22:53:14Z", + "result_id": "node_9a8e7c2231a3", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Native fixed50: v26 49554485611 (50/50), v35 49596964560 (50/50). v35 is the measured incumbent; controlled optimization continues.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "e076742e1f3fecf3", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T22:55:15+00:00", + "result_id": "node_9a8e7c2231a3", + "run_id": "explore-result:2af069d82cacd923a4060e00cb5731824852192e4e2bb50d8c45f5a4b43e798f", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Native fixed50: v26 49554485611 (50/50), v35 49596964560 (50/50). v35 is the measured incumbent; controlled optimization continues.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "99ab695db4ca4b4b", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-uppertemp-native" + ], + "execution_observation": { + "candidate_sha256": "8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "evidence:ad-uppertemp-native" + ], + "executed_method_id": "ad-placement.native.fixed50.v1", + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "limit_seconds": null, + "planned_method_id": "ad-placement.native.fixed50.v1", + "result_candidate_sha256": "8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "result_ref": "evidence:ad-uppertemp-native", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Accepted native evaluation completed; resource details are not exposed." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:native:current-task:fixed50", + "evaluation_ref": "native-protocol:README-sha256:0051d2a6e88e299efc0c0b0f1a7e8a68b3060c855b55ffce8c6cfa601d4b30b2", + "evidence_ref": "evidence:ad-uppertemp-native", + "higher_is_better": true, + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "metric": "total_native_score", + "unit": "points_out_of_50e9", + "value": 49596964560 + } + ], + "node_id": "node_9a8e7c2231a3", + "recorded_at": "2026-10-22T22:55:15+00:00", + "result_id": "finding_2af069d82cacd923a4060e00cb5731824852192e4e2bb50d8c45f5a4b43e798f", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50 Applicability: Native fixed50: v26 49554485611 (50/50), v35 49596964560 (50/50). v35 is the measured incumbent; controlled optimization continues. Observation: Frozen v35 scored 49596964560 with 50/50 passes in submission 58b3cc50d0d5. Delta versus v26: 42478949. Retained v35. Interpretation: The actual accepted ID and result receipt match the frozen source archive. The native score improved and the trial source is retained. Local candidate selection did not certify native gain. Public per-test statuses are preserved; per-case scores, execution time, memory and iterations were not exposed. Null resource fields mean unmeasured. P1 remains open. Next is the separate, unimplemented mixed-epoch temperature probe.", + "tags": [ + "writeback-result" + ], + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "0233ae0520cc0746", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v35-mixedtemp-train/scores.tsv", + "research/wake12-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:mixedtemp-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v35-mixedtemp-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992589538.2 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T23:13:34Z", + "result_id": "finding_33c740d49874", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 source8af4a746 averaged992589538.2 on20 original-score/O2/4.70-second generated cases. Scope:train. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Mixed-temperature v35 train mean 992589538.2" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "8ef1979a1e7fc742", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v37-mixedtemp-train/scores.tsv", + "research/wake12-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:de576f0752656368404f1ae091e8022e23e4e675d3bef52cbeac1c1ab4e89eb3", + "comparison_scope": "ad-placement:local:mixedtemp-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v37-mixedtemp-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992627654 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T23:13:38Z", + "result_id": "finding_9dfeb7b81500", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v37 sourcede576f07 averaged992627654 on20 original-score/O2/4.70-second generated cases. Scope:train. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Mixed-temperature v37 train mean 992627654" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "f87d1c8aef434c6b", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v38-mixedtemp-train/scores.tsv", + "research/wake12-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:30b0ee611cb72690834a4c864899367df10d29dc102c6ab12eb3df08cdf1d4e5", + "comparison_scope": "ad-placement:local:mixedtemp-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v38-mixedtemp-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992489008.85 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T23:13:40Z", + "result_id": "finding_d378a8a5d746", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v38 source30b0ee61 averaged992489008.85 on20 original-score/O2/4.70-second generated cases. Scope:train. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Mixed-temperature v38 train mean 992489008.85" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "a21233d3952f12b2", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v35-mixedtemp-heldout/scores.tsv", + "research/wake12-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:mixedtemp-heldout:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v35-mixedtemp-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:8157adca969ac3551ea8c3c22c3429c2aa82610340a2a77e3a0b8cadc4a71489", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992038467.1 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T23:13:42Z", + "result_id": "finding_dc1b44087ab3", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 source8af4a746 averaged992038467.1 on20 original-score/O2/4.70-second generated cases. Scope:heldout. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Mixed-temperature v35 heldout mean 992038467.1" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "e571e484c5e43815", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v37-mixedtemp-heldout/scores.tsv", + "research/wake12-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:de576f0752656368404f1ae091e8022e23e4e675d3bef52cbeac1c1ab4e89eb3", + "comparison_scope": "ad-placement:local:mixedtemp-heldout:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v37-mixedtemp-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:8157adca969ac3551ea8c3c22c3429c2aa82610340a2a77e3a0b8cadc4a71489", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991719062.85 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T23:13:44Z", + "result_id": "finding_9e96b1c7eb40", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v37 sourcede576f07 averaged991719062.85 on20 original-score/O2/4.70-second generated cases. Scope:heldout. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Mixed-temperature v37 heldout mean 991719062.85" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "823948bcee2d4277", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v35-mixedtemp-extra/scores.tsv", + "research/wake12-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:mixedtemp-extra:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v35-mixedtemp-extra/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:5fee0a07fc90dc346fbd5aef792cd64262f5b0022472a73036073cd9f702b295", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992550870.85 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T23:13:46Z", + "result_id": "finding_9be5503ff78f", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 source8af4a746 averaged992550870.85 on20 original-score/O2/4.70-second generated cases. Scope:extra. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Mixed-temperature v35 extra mean 992550870.85" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "e99c984d73fb50bc", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v37-mixedtemp-extra/scores.tsv", + "research/wake12-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:de576f0752656368404f1ae091e8022e23e4e675d3bef52cbeac1c1ab4e89eb3", + "comparison_scope": "ad-placement:local:mixedtemp-extra:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v37-mixedtemp-extra/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:5fee0a07fc90dc346fbd5aef792cd64262f5b0022472a73036073cd9f702b295", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991011262.05 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T23:13:48Z", + "result_id": "finding_2c3bcb1aa5ef", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v37 sourcede576f07 averaged991011262.05 on20 original-score/O2/4.70-second generated cases. Scope:extra. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Mixed-temperature v37 extra mean 991011262.05" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "18e47cade2d04531", + "event_kind": "finding", + "evidence_refs": [ + "research/mixedtemp-fresh40-summary.json", + "research/wake12-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:mixedtemp-fresh40:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/mixedtemp-fresh40-summary.json", + "higher_is_better": true, + "input_revision": "sha256:1ce94581646353ee46cb17ef88b273ee8b9a07bca96fb92585c8c368df6e0e71", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992294668.975 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T23:13:50Z", + "result_id": "finding_316d172ed654", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 source8af4a746 averaged992294668.975 on40 original-score/O2/4.70-second generated cases. Scope:fresh40. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Mixed-temperature v35 fresh40 mean 992294668.975" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "b6a36ecb9350d485", + "event_kind": "finding", + "evidence_refs": [ + "research/mixedtemp-fresh40-summary.json", + "research/wake12-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:de576f0752656368404f1ae091e8022e23e4e675d3bef52cbeac1c1ab4e89eb3", + "comparison_scope": "ad-placement:local:mixedtemp-fresh40:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/mixedtemp-fresh40-summary.json", + "higher_is_better": true, + "input_revision": "sha256:1ce94581646353ee46cb17ef88b273ee8b9a07bca96fb92585c8c368df6e0e71", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991365162.45 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T23:13:52Z", + "result_id": "finding_a4dd87ba39bf", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v37 sourcede576f07 averaged991365162.45 on40 original-score/O2/4.70-second generated cases. Scope:fresh40. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Mixed-temperature v37 fresh40 mean 991365162.45" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "8c6297e6b2952101", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T23:14:09Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 tests. v37 lost 929507 mean points on40 fresh cases (12wins); uncertainty retained. Not adopted; nativev35 retained.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "147800cca3589cba", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T23:15:40+00:00", + "result_id": "node_e388daf41666", + "run_id": "explore-result:25056191a7aee2593df53a68f77a8af6fd0966d80e72cae80d40d3d8732aebfc", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 tests. v37 lost 929507 mean points on40 fresh cases (12wins); uncertainty retained. Not adopted; nativev35 retained.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "5fab424d7a841736", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-mixedtemp-fresh40" + ], + "execution_observation": { + "candidate_sha256": "de576f0752656368404f1ae091e8022e23e4e675d3bef52cbeac1c1ab4e89eb3", + "change_reason": null, + "elapsed_seconds": 189.16512999999998, + "evidence_refs": [ + "evidence:ad-mixedtemp-fresh40" + ], + "executed_method_id": "ad-placement.paired.40x4p70s.v1", + "exit_code": 0, + "input_revision": "sha256:1ce94581646353ee46cb17ef88b273ee8b9a07bca96fb92585c8c368df6e0e71", + "limit_seconds": 200, + "planned_method_id": "ad-placement.paired.40x4p70s.v1", + "result_candidate_sha256": "de576f0752656368404f1ae091e8022e23e4e675d3bef52cbeac1c1ab4e89eb3", + "result_ref": "evidence:ad-mixedtemp-fresh40", + "schema_version": "explore_execution_observation_v0", + "search_iterations": 284313384, + "stop_reason": "All40 serial searches completed within5seconds each; both predeclared cohorts retained." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:de576f0752656368404f1ae091e8022e23e4e675d3bef52cbeac1c1ab4e89eb3", + "comparison_scope": "ad-placement:local:mixedtemp-fresh40:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "evidence:ad-mixedtemp-fresh40", + "higher_is_better": true, + "input_revision": "sha256:1ce94581646353ee46cb17ef88b273ee8b9a07bca96fb92585c8c368df6e0e71", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991365162.45 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T23:15:40+00:00", + "result_id": "finding_25056191a7aee2593df53a68f77a8af6fd0966d80e72cae80d40d3d8732aebfc", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:1ce94581646353ee46cb17ef88b273ee8b9a07bca96fb92585c8c368df6e0e71 Applicability: 4.70-second O2 tests. v37 lost 929507 mean points on40 fresh cases (12wins); uncertainty retained. Not adopted; nativev35 retained. Observation: v37 averaged 991365162.45 on40 fresh cases, delta -929506.525 versus v35, 12/40 wins. All checks passed; maximum 4.7577s. Native result is unmeasured. Interpretation: The mixed order was frozen on training before either fresh cohort was generated or scored. Both cohorts are retained. Exploratory95% paired-case interval [-1517372.525,-363056.975] omits scheduling noise, selection multiplicity and native shift. The tested mixed order is not adopted; this does not refute all mixed schedules. Diagnostics verify reached temperature paths, not causal quality effects. Execution fields aggregate40 cases. The next geometry probe is separate and unimplemented.", + "tags": [ + "writeback-result" + ], + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "307b13f4478077a4", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v35-aspectblend-train/scores.tsv", + "research/wake13-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:aspectblend-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v35-aspectblend-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992192528.15 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T23:36:05Z", + "result_id": "finding_7c1cffd5ac4e", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 source8af4a746 averaged992192528.15 on20 original-score/O2/4.70-second generated cases. Scope:train. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Global-aspect target v35 train mean 992192528.15" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "e9bdf1b357162c56", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v39-aspectblend-train/scores.tsv", + "research/wake13-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:f9fb35eb6494295c4e153abd54ad24d3d69f04888cbda414e04dd9c8468e95ac", + "comparison_scope": "ad-placement:local:aspectblend-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v39-aspectblend-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992824988.35 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T23:36:08Z", + "result_id": "finding_8e2321e5abcc", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v39 sourcef9fb35eb averaged992824988.35 on20 original-score/O2/4.70-second generated cases. Scope:train. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Global-aspect target v39 train mean 992824988.35" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "a8ac9c8cd91affdf", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v40-aspectblend-train/scores.tsv", + "research/wake13-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:4867af10a078ff467a0f915b0bbc10032201d668ac525a7758d2dca7158bbf39", + "comparison_scope": "ad-placement:local:aspectblend-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v40-aspectblend-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992479972.95 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T23:36:09Z", + "result_id": "finding_024797399183", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v40 source4867af10 averaged992479972.95 on20 original-score/O2/4.70-second generated cases. Scope:train. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Global-aspect target v40 train mean 992479972.95" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "9665723d0116da96", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v35-aspectblend-heldout/scores.tsv", + "research/wake13-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:aspectblend-heldout:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v35-aspectblend-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:7fe3991c0f85ff27884d85fce559ef6a5bf037aae92239506114ea19676146ee", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991044103.05 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T23:36:11Z", + "result_id": "finding_6a7c9ffabb34", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 source8af4a746 averaged991044103.05 on20 original-score/O2/4.70-second generated cases. Scope:heldout. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Global-aspect target v35 heldout mean 991044103.05" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "195e7eedd723b85f", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v39-aspectblend-heldout/scores.tsv", + "research/wake13-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:f9fb35eb6494295c4e153abd54ad24d3d69f04888cbda414e04dd9c8468e95ac", + "comparison_scope": "ad-placement:local:aspectblend-heldout:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v39-aspectblend-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:7fe3991c0f85ff27884d85fce559ef6a5bf037aae92239506114ea19676146ee", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991330812.9 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T23:36:12Z", + "result_id": "finding_0df128d32b8f", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v39 sourcef9fb35eb averaged991330812.9 on20 original-score/O2/4.70-second generated cases. Scope:heldout. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Global-aspect target v39 heldout mean 991330812.9" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "c259220c8548f829", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v35-aspectblend-extra/scores.tsv", + "research/wake13-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:aspectblend-extra:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v35-aspectblend-extra/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:4d540da60ace460d78cc0e8a1f8093918c6b9221836e2026fb91b7478d88c214", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990203599.25 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T23:36:14Z", + "result_id": "finding_08274bbe5edf", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 source8af4a746 averaged990203599.25 on20 original-score/O2/4.70-second generated cases. Scope:extra. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Global-aspect target v35 extra mean 990203599.25" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "f8b042e82e3b4208", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v39-aspectblend-extra/scores.tsv", + "research/wake13-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:f9fb35eb6494295c4e153abd54ad24d3d69f04888cbda414e04dd9c8468e95ac", + "comparison_scope": "ad-placement:local:aspectblend-extra:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v39-aspectblend-extra/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:4d540da60ace460d78cc0e8a1f8093918c6b9221836e2026fb91b7478d88c214", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990655910.45 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T23:36:15Z", + "result_id": "finding_06bc356bcac1", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v39 sourcef9fb35eb averaged990655910.45 on20 original-score/O2/4.70-second generated cases. Scope:extra. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Global-aspect target v39 extra mean 990655910.45" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "5027e51fc7ae26a2", + "event_kind": "finding", + "evidence_refs": [ + "research/aspectblend-fresh40-summary.json", + "research/wake13-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:aspectblend-fresh40:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/aspectblend-fresh40-summary.json", + "higher_is_better": true, + "input_revision": "sha256:bd097bc49a8107f295f2f1cfcf0225a48f2a5a0f291b7fd604ea1a2f0af0564b", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990623851.15 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T23:36:16Z", + "result_id": "finding_cf27c524b19a", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 source8af4a746 averaged990623851.15 on40 original-score/O2/4.70-second generated cases. Scope:fresh40. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Global-aspect target v35 fresh40 mean 990623851.15" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "28ca5fd984596960", + "event_kind": "finding", + "evidence_refs": [ + "research/aspectblend-fresh40-summary.json", + "research/wake13-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:f9fb35eb6494295c4e153abd54ad24d3d69f04888cbda414e04dd9c8468e95ac", + "comparison_scope": "ad-placement:local:aspectblend-fresh40:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/aspectblend-fresh40-summary.json", + "higher_is_better": true, + "input_revision": "sha256:bd097bc49a8107f295f2f1cfcf0225a48f2a5a0f291b7fd604ea1a2f0af0564b", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990993361.675 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T23:36:18Z", + "result_id": "finding_96ef2a40ab9d", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v39 sourcef9fb35eb averaged990993361.675 on40 original-score/O2/4.70-second generated cases. Scope:fresh40. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Global-aspect target v39 fresh40 mean 990993361.675" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "8090f719f279f6c7", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T23:36:52Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 tests. v39 gained 369511 mean points on40 fresh cases (22wins); uncertainty retained. Pending native; nativev35 retained.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "61c2612063753351", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-22T23:38:24+00:00", + "result_id": "node_e388daf41666", + "run_id": "explore-result:fdea7c5ae0be3f07fdf51c4ddc39f90307e2955a461ce9c9175ea3c034fa5d95", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 tests. v39 gained 369511 mean points on40 fresh cases (22wins); uncertainty retained. Pending native; nativev35 retained.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "dc3a1ff7f707f632", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-aspectblend-fresh40" + ], + "execution_observation": { + "candidate_sha256": "f9fb35eb6494295c4e153abd54ad24d3d69f04888cbda414e04dd9c8468e95ac", + "change_reason": null, + "elapsed_seconds": 189.10896999999994, + "evidence_refs": [ + "evidence:ad-aspectblend-fresh40" + ], + "executed_method_id": "ad-placement.paired.40x4p70s.v1", + "exit_code": 0, + "input_revision": "sha256:bd097bc49a8107f295f2f1cfcf0225a48f2a5a0f291b7fd604ea1a2f0af0564b", + "limit_seconds": 200, + "planned_method_id": "ad-placement.paired.40x4p70s.v1", + "result_candidate_sha256": "f9fb35eb6494295c4e153abd54ad24d3d69f04888cbda414e04dd9c8468e95ac", + "result_ref": "evidence:ad-aspectblend-fresh40", + "schema_version": "explore_execution_observation_v0", + "search_iterations": 261915432, + "stop_reason": "All40 serial searches completed within5seconds each; both predeclared cohorts retained." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:f9fb35eb6494295c4e153abd54ad24d3d69f04888cbda414e04dd9c8468e95ac", + "comparison_scope": "ad-placement:local:aspectblend-fresh40:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "evidence:ad-aspectblend-fresh40", + "higher_is_better": true, + "input_revision": "sha256:bd097bc49a8107f295f2f1cfcf0225a48f2a5a0f291b7fd604ea1a2f0af0564b", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 990993361.675 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-22T23:38:24+00:00", + "result_id": "finding_fdea7c5ae0be3f07fdf51c4ddc39f90307e2955a461ce9c9175ea3c034fa5d95", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:bd097bc49a8107f295f2f1cfcf0225a48f2a5a0f291b7fd604ea1a2f0af0564b Applicability: 4.70-second O2 tests. v39 gained 369511 mean points on40 fresh cases (22wins); uncertainty retained. Pending native; nativev35 retained. Observation: v39 averaged 990993361.675 on40 fresh cases, delta 369510.525 versus v35, 22/40 wins. All checks passed; maximum 4.75596s. Native result is unmeasured. Interpretation: The selected fraction was frozen on training before generating/scoring either fresh cohort. Both cohorts are retained. Exploratory95% paired-case interval [-133062.55,864689.25] omits scheduling noise, selection multiplicity and native shift. The positive mean supports a future source-matched native trial; exact v35 remains retained. A stale diagnostic executable name and missing counter hook were fixed before quality scoring; the original attempt is archived and candidate sources did not change. Corrected diagnostics verify target expansion, not causal quality effects. Execution fields aggregate40 cases. The next probe is separate and unimplemented.", + "tags": [ + "writeback-result" + ], + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "36d9f34a3a58c97f", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v35-stepdecay-train/scores.tsv", + "research/wake14-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:stepdecay-train:gxx-O2:4.70s:roundrobin3-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v35-stepdecay-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992196140.2 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:01:14Z", + "result_id": "finding_f08e9c829372", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 source8af4a746 averaged992196140.2 on20 original-score/O2/4.70-second generated cases. Scope:train. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Proposal-step study v35 train mean 992196140.2" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "e3ba8e44037ff4bf", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v41-stepdecay-train/scores.tsv", + "research/wake14-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:cf1d31cff75ed16a79796604e28520339b54b50ffd0ef9eae663bc61bbacbe16", + "comparison_scope": "ad-placement:local:stepdecay-train:gxx-O2:4.70s:roundrobin3-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v41-stepdecay-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992021709.75 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:01:15Z", + "result_id": "finding_cd84ea0f1db0", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v41 sourcecf1d31cf averaged992021709.75 on20 original-score/O2/4.70-second generated cases. Scope:train. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Proposal-step study v41 train mean 992021709.75" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "d4f1694c49cb4488", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v42-stepdecay-train/scores.tsv", + "research/wake14-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:29e3dfc8b0ef28e9fa4a923d076de839b28580020659053a697647406b0c30bc", + "comparison_scope": "ad-placement:local:stepdecay-train:gxx-O2:4.70s:roundrobin3-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v42-stepdecay-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992517910.3 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:01:16Z", + "result_id": "finding_8bc36b1be357", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v42 source29e3dfc8 averaged992517910.3 on20 original-score/O2/4.70-second generated cases. Scope:train. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Proposal-step study v42 train mean 992517910.3" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "49d6c96422ad5ad5", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v35-stepdecay-heldout/scores.tsv", + "research/wake14-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:stepdecay-heldout:gxx-O2:4.70s:roundrobin3-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v35-stepdecay-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:c441ee887fc27dbe99846f7c6deeff50803412a8a7900ef00224ea7551118628", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992082557.1 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:01:18Z", + "result_id": "finding_8568cf5615dc", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 source8af4a746 averaged992082557.1 on20 original-score/O2/4.70-second generated cases. Scope:heldout. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Proposal-step study v35 heldout mean 992082557.1" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "0c4ba5cf768442d1", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v39-stepdecay-heldout/scores.tsv", + "research/wake14-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:f9fb35eb6494295c4e153abd54ad24d3d69f04888cbda414e04dd9c8468e95ac", + "comparison_scope": "ad-placement:local:stepdecay-heldout:gxx-O2:4.70s:roundrobin3-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v39-stepdecay-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:c441ee887fc27dbe99846f7c6deeff50803412a8a7900ef00224ea7551118628", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992709448.9 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:01:19Z", + "result_id": "finding_0fa0bcae666b", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v39 sourcef9fb35eb averaged992709448.9 on20 original-score/O2/4.70-second generated cases. Scope:heldout. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Proposal-step study v39 heldout mean 992709448.9" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "21c600c743cf9c23", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v42-stepdecay-heldout/scores.tsv", + "research/wake14-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:29e3dfc8b0ef28e9fa4a923d076de839b28580020659053a697647406b0c30bc", + "comparison_scope": "ad-placement:local:stepdecay-heldout:gxx-O2:4.70s:roundrobin3-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v42-stepdecay-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:c441ee887fc27dbe99846f7c6deeff50803412a8a7900ef00224ea7551118628", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991787100.1 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:01:20Z", + "result_id": "finding_cc9282ab8b56", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v42 source29e3dfc8 averaged991787100.1 on20 original-score/O2/4.70-second generated cases. Scope:heldout. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Proposal-step study v42 heldout mean 991787100.1" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "212d21d384e2cd8d", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v35-stepdecay-extra/scores.tsv", + "research/wake14-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:stepdecay-extra:gxx-O2:4.70s:roundrobin3-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v35-stepdecay-extra/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:db0ed91165e30d3ff6aeb4f0c3ae5626d02cf1eb490fe3ed4da11e9897dd8474", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992143904.5 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:01:21Z", + "result_id": "finding_f63b50f1a52e", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 source8af4a746 averaged992143904.5 on20 original-score/O2/4.70-second generated cases. Scope:extra. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Proposal-step study v35 extra mean 992143904.5" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "9f301ad8b9301721", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v39-stepdecay-extra/scores.tsv", + "research/wake14-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:f9fb35eb6494295c4e153abd54ad24d3d69f04888cbda414e04dd9c8468e95ac", + "comparison_scope": "ad-placement:local:stepdecay-extra:gxx-O2:4.70s:roundrobin3-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v39-stepdecay-extra/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:db0ed91165e30d3ff6aeb4f0c3ae5626d02cf1eb490fe3ed4da11e9897dd8474", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992170174.05 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:01:23Z", + "result_id": "finding_88a91f17375a", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v39 sourcef9fb35eb averaged992170174.05 on20 original-score/O2/4.70-second generated cases. Scope:extra. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Proposal-step study v39 extra mean 992170174.05" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "8cd5291b335cb8ba", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v42-stepdecay-extra/scores.tsv", + "research/wake14-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:29e3dfc8b0ef28e9fa4a923d076de839b28580020659053a697647406b0c30bc", + "comparison_scope": "ad-placement:local:stepdecay-extra:gxx-O2:4.70s:roundrobin3-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v42-stepdecay-extra/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:db0ed91165e30d3ff6aeb4f0c3ae5626d02cf1eb490fe3ed4da11e9897dd8474", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992486299.3 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:01:24Z", + "result_id": "finding_ce11891529b3", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v42 source29e3dfc8 averaged992486299.3 on20 original-score/O2/4.70-second generated cases. Scope:extra. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Proposal-step study v42 extra mean 992486299.3" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "180be0546bc2e922", + "event_kind": "finding", + "evidence_refs": [ + "research/stepdecay-fresh40-summary.json", + "research/wake14-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:stepdecay-fresh40:gxx-O2:4.70s:roundrobin3-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/stepdecay-fresh40-summary.json", + "higher_is_better": true, + "input_revision": "sha256:3d5679e3bbddb21e007be972472c585a687094eda42ed7c8aa31bc1ac3da5dd9", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992113230.8 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:01:25Z", + "result_id": "finding_718715acf801", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 source8af4a746 averaged992113230.8 on40 original-score/O2/4.70-second generated cases. Scope:fresh40. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Proposal-step study v35 fresh40 mean 992113230.8" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "d7b7f2d1cd13675a", + "event_kind": "finding", + "evidence_refs": [ + "research/stepdecay-fresh40-summary.json", + "research/wake14-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:f9fb35eb6494295c4e153abd54ad24d3d69f04888cbda414e04dd9c8468e95ac", + "comparison_scope": "ad-placement:local:stepdecay-fresh40:gxx-O2:4.70s:roundrobin3-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/stepdecay-fresh40-summary.json", + "higher_is_better": true, + "input_revision": "sha256:3d5679e3bbddb21e007be972472c585a687094eda42ed7c8aa31bc1ac3da5dd9", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992439811.475 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:01:27Z", + "result_id": "finding_ab74af0a0a4c", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v39 sourcef9fb35eb averaged992439811.475 on40 original-score/O2/4.70-second generated cases. Scope:fresh40. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Proposal-step study v39 fresh40 mean 992439811.475" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "a41b608fce93f4e5", + "event_kind": "finding", + "evidence_refs": [ + "research/stepdecay-fresh40-summary.json", + "research/wake14-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:29e3dfc8b0ef28e9fa4a923d076de839b28580020659053a697647406b0c30bc", + "comparison_scope": "ad-placement:local:stepdecay-fresh40:gxx-O2:4.70s:roundrobin3-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/stepdecay-fresh40-summary.json", + "higher_is_better": true, + "input_revision": "sha256:3d5679e3bbddb21e007be972472c585a687094eda42ed7c8aa31bc1ac3da5dd9", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992136699.7 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:01:28Z", + "result_id": "finding_0d3347ca4bc1", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v42 source29e3dfc8 averaged992136699.7 on40 original-score/O2/4.70-second generated cases. Scope:fresh40. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Proposal-step study v42 fresh40 mean 992136699.7" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "f0375388791f0f03", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T00:03:57Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 tests. v39 selected on40 common fresh cases, delta326581 vs v35. Selection uncertainty retained. Native incumbent v35.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "97bd3c62fd151199", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T00:03:58Z", + "result_id": "node_9a8e7c2231a3", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Native fixed50: v35 49596964560 (50/50), v39 49590533157 (50/50). v35 is the measured incumbent; controlled optimization continues.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "b85e3f31ee750e28", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T00:05:52+00:00", + "result_id": "node_9a8e7c2231a3", + "run_id": "explore-result:c1ea83ba19b42e785cf0436a7ab77335159f2ef0fa61544aede9234b1c7be9d7", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Native fixed50: v35 49596964560 (50/50), v39 49590533157 (50/50). v35 is the measured incumbent; controlled optimization continues.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "5db97a692cd3c0c3", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-stepdecay-native" + ], + "execution_observation": { + "candidate_sha256": "f9fb35eb6494295c4e153abd54ad24d3d69f04888cbda414e04dd9c8468e95ac", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "evidence:ad-stepdecay-native" + ], + "executed_method_id": "ad-placement.native.fixed50.v1", + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "limit_seconds": null, + "planned_method_id": "ad-placement.native.fixed50.v1", + "result_candidate_sha256": "f9fb35eb6494295c4e153abd54ad24d3d69f04888cbda414e04dd9c8468e95ac", + "result_ref": "evidence:ad-stepdecay-native", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Accepted native evaluation completed; resource details not exposed." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:f9fb35eb6494295c4e153abd54ad24d3d69f04888cbda414e04dd9c8468e95ac", + "comparison_scope": "ad-placement:native:current-task:fixed50", + "evaluation_ref": "native-protocol:README-sha256:0051d2a6e88e299efc0c0b0f1a7e8a68b3060c855b55ffce8c6cfa601d4b30b2", + "evidence_ref": "evidence:ad-stepdecay-native", + "higher_is_better": true, + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "metric": "total_native_score", + "unit": "points_out_of_50e9", + "value": 49590533157 + } + ], + "node_id": "node_9a8e7c2231a3", + "recorded_at": "2026-10-23T00:05:52+00:00", + "result_id": "finding_c1ea83ba19b42e785cf0436a7ab77335159f2ef0fa61544aede9234b1c7be9d7", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50 Applicability: Native fixed50: v35 49596964560 (50/50), v39 49590533157 (50/50). v35 is the measured incumbent; controlled optimization continues. Observation: Frozen v39 scored 49,590,533,157 with 50/50 passes in submission 04e240d34f3a. Delta versus v35: -6,431,403. Retained v35. Interpretation: The actual accepted ID and result receipt match the frozen source archive. Trial did not improve the valid incumbent; exact v35 restored. Local selection did not certify native gain. Public case statuses are preserved; individual scores, execution time, memory and iterations were not exposed. Null resources mean unmeasured. The step training winner was frozen before both reserved cohorts; native source ranking then used all 40 common fresh cases and is selection, not independent confirmation. P1 remains open. Next is the separate, unimplemented local/global aspect-frequency probe.", + "tags": [ + "writeback-result" + ], + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "e87c7f0681263f0e", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v35-aspectfreq-train/scores.tsv", + "research/wake15-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:aspectfreq-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v35-aspectfreq-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992544686.75 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:23:56Z", + "result_id": "finding_0bdf1c48d243", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 source8af4a746 averaged992544686.75 on20 original-score/O2/4.70-second generated cases. Scope:train. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Aspect proposal frequency v35 train mean 992544686.75" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "a124dda0c8774de9", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v43-aspectfreq-train/scores.tsv", + "research/wake15-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:ba4322a5a7e436ccdc5092c561419b48e3cde803d50f86789beeca073b6be5fc", + "comparison_scope": "ad-placement:local:aspectfreq-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v43-aspectfreq-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992794650.95 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:23:59Z", + "result_id": "finding_2b06d8298d31", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v43 sourceba4322a5 averaged992794650.95 on20 original-score/O2/4.70-second generated cases. Scope:train. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Aspect proposal frequency v43 train mean 992794650.95" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "6d7c0191798850d2", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v44-aspectfreq-train/scores.tsv", + "research/wake15-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:1d9b61cd27503d2f046a0767dfe5c37527946dc2626f55b81c86af952799c6f1", + "comparison_scope": "ad-placement:local:aspectfreq-train:gxx-O2:4.70s:roundrobin-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v44-aspectfreq-train/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:99e141ac1325c46ce34eec62d2edd503524a840a23e933e83e094cdf9475d05d", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992525935.95 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:24:00Z", + "result_id": "finding_e12a3a4ee4ab", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v44 source1d9b61cd averaged992525935.95 on20 original-score/O2/4.70-second generated cases. Scope:train. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Aspect proposal frequency v44 train mean 992525935.95" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "990002aac8af2e83", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v35-aspectfreq-heldout/scores.tsv", + "research/wake15-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:aspectfreq-heldout:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v35-aspectfreq-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:e8a5d627e1088271f4f87c51cffa78808ffdc0bbd33f07b8ec04d7fbfd5c82b9", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992861526.9 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:24:01Z", + "result_id": "finding_1eb8151ef312", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 source8af4a746 averaged992861526.9 on20 original-score/O2/4.70-second generated cases. Scope:heldout. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Aspect proposal frequency v35 heldout mean 992861526.9" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "5e5ef5107bef0b57", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v43-aspectfreq-heldout/scores.tsv", + "research/wake15-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:ba4322a5a7e436ccdc5092c561419b48e3cde803d50f86789beeca073b6be5fc", + "comparison_scope": "ad-placement:local:aspectfreq-heldout:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v43-aspectfreq-heldout/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:e8a5d627e1088271f4f87c51cffa78808ffdc0bbd33f07b8ec04d7fbfd5c82b9", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992825695.6 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:24:03Z", + "result_id": "finding_80d6fd119da9", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v43 sourceba4322a5 averaged992825695.6 on20 original-score/O2/4.70-second generated cases. Scope:heldout. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Aspect proposal frequency v43 heldout mean 992825695.6" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "07954717a8c51c76", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v35-aspectfreq-extra/scores.tsv", + "research/wake15-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:aspectfreq-extra:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v35-aspectfreq-extra/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:dad1dd8d600a66385b5d083c74349aa5afe12e4138d159aa6686588832a31df1", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991561712.9 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:24:04Z", + "result_id": "finding_1d97feee19d3", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 source8af4a746 averaged991561712.9 on20 original-score/O2/4.70-second generated cases. Scope:extra. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Aspect proposal frequency v35 extra mean 991561712.9" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "622b6ef2c560f0b0", + "event_kind": "finding", + "evidence_refs": [ + "research/results/v43-aspectfreq-extra/scores.tsv", + "research/wake15-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:ba4322a5a7e436ccdc5092c561419b48e3cde803d50f86789beeca073b6be5fc", + "comparison_scope": "ad-placement:local:aspectfreq-extra:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/results/v43-aspectfreq-extra/scores.tsv", + "higher_is_better": true, + "input_revision": "sha256:dad1dd8d600a66385b5d083c74349aa5afe12e4138d159aa6686588832a31df1", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991831264.15 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:24:05Z", + "result_id": "finding_ff66862a7694", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v43 sourceba4322a5 averaged991831264.15 on20 original-score/O2/4.70-second generated cases. Scope:extra. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Aspect proposal frequency v43 extra mean 991831264.15" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "4552d161a2ea35f3", + "event_kind": "finding", + "evidence_refs": [ + "research/aspectfreq-fresh40-summary.json", + "research/wake15-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:aspectfreq-fresh40:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/aspectfreq-fresh40-summary.json", + "higher_is_better": true, + "input_revision": "sha256:b7029e4ad83baa0ba1fccd7538e6a78db803f2361380bfeada10c1f80aea34c1", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992211619.9 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:24:06Z", + "result_id": "finding_9ef58f0f8ede", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 source8af4a746 averaged992211619.9 on40 original-score/O2/4.70-second generated cases. Scope:fresh40. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Aspect proposal frequency v35 fresh40 mean 992211619.9" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "900cac74ca8e9b6a", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T00:24:06Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 tests. v43 gained 116860 mean points on40 fresh cases (19wins); uncertainty retained. Pending native; native v35 retained.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "a0d56a1dd5f5d192", + "event_kind": "finding", + "evidence_refs": [ + "research/aspectfreq-fresh40-summary.json", + "research/wake15-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:ba4322a5a7e436ccdc5092c561419b48e3cde803d50f86789beeca073b6be5fc", + "comparison_scope": "ad-placement:local:aspectfreq-fresh40:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/aspectfreq-fresh40-summary.json", + "higher_is_better": true, + "input_revision": "sha256:b7029e4ad83baa0ba1fccd7538e6a78db803f2361380bfeada10c1f80aea34c1", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992328479.875 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:24:07Z", + "result_id": "finding_a98e2874bafb", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v43 sourceba4322a5 averaged992328479.875 on40 original-score/O2/4.70-second generated cases. Scope:fresh40. All tester and independent geometry checks passed. This local comparison does not establish a native gain.", + "title": "Aspect proposal frequency v43 fresh40 mean 992328479.875" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "1f0f34c1a2e7eaab", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T00:26:00+00:00", + "result_id": "node_e388daf41666", + "run_id": "explore-result:3fabd8418fff49b2371878ffaf56b98bffb9543e2e3a5e04ed5382a3067b8b46", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 tests. v43 gained 116860 mean points on40 fresh cases (19wins); uncertainty retained. Pending native; native v35 retained.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "b461f91109ff9300", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-aspectfreq-fresh40" + ], + "execution_observation": { + "candidate_sha256": "ba4322a5a7e436ccdc5092c561419b48e3cde803d50f86789beeca073b6be5fc", + "change_reason": null, + "elapsed_seconds": 188.8788, + "evidence_refs": [ + "evidence:ad-aspectfreq-fresh40" + ], + "executed_method_id": "ad-placement.paired.40x4p70s.v1", + "exit_code": 0, + "input_revision": "sha256:b7029e4ad83baa0ba1fccd7538e6a78db803f2361380bfeada10c1f80aea34c1", + "limit_seconds": 200, + "planned_method_id": "ad-placement.paired.40x4p70s.v1", + "result_candidate_sha256": "ba4322a5a7e436ccdc5092c561419b48e3cde803d50f86789beeca073b6be5fc", + "result_ref": "evidence:ad-aspectfreq-fresh40", + "schema_version": "explore_execution_observation_v0", + "search_iterations": 351224360, + "stop_reason": "All 40 serial searches completed within 5 seconds each; both predeclared cohorts retained." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:ba4322a5a7e436ccdc5092c561419b48e3cde803d50f86789beeca073b6be5fc", + "comparison_scope": "ad-placement:local:aspectfreq-fresh40:gxx-O2:4.70s:paired-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "evidence:ad-aspectfreq-fresh40", + "higher_is_better": true, + "input_revision": "sha256:b7029e4ad83baa0ba1fccd7538e6a78db803f2361380bfeada10c1f80aea34c1", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992328479.875 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:26:00+00:00", + "result_id": "finding_3fabd8418fff49b2371878ffaf56b98bffb9543e2e3a5e04ed5382a3067b8b46", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:b7029e4ad83baa0ba1fccd7538e6a78db803f2361380bfeada10c1f80aea34c1 Applicability: 4.70-second O2 tests. v43 gained 116860 mean points on40 fresh cases (19wins); uncertainty retained. Pending native; native v35 retained. Observation: v43 averaged 992328479.875 on 40 fresh cases, delta 116859.975 versus v35, 19/40 wins. All checks passed; maximum 4.73404s. Native result is unmeasured. Interpretation: The selected frequency was frozen on training before generating or scoring either fresh cohort. Both complete cohorts are retained. Exploratory 95% paired-case interval [-232679.875,475818.55] omits scheduling noise, selection multiplicity and native shift. The positive mean supports a future native trial; exact v35 stays retained. Instrumented diagnostics verify executed branch ranges and attempt frequencies, not causal quality effects. Execution fields aggregate 40 cases. The completed v39 regression and positive local evidence remain reconciled; native details do not identify its cause. The next probe is separate and unimplemented.", + "tags": [ + "writeback-result" + ], + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "c18c2fc87dd3837d", + "event_kind": "finding", + "evidence_refs": [ + "research/v43-confirmation-summary.json", + "research/wake17-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:v43-confirmation-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/v43-confirmation-summary.json", + "higher_is_better": true, + "input_revision": "sha256:67c3352c9f08cf20fc4aaa04abfaa0dd74d91c34f556f8c8effd5dce2f370b66", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 993447588.8 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:57:09Z", + "result_id": "finding_ad6e72381f77", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 averaged 993447588.8 over40 new cases with two4.70-second searches per case. ABBA/BAAB order and case-group analysis preserve repeated observations. All80 source runs passed tester, geometry and resources. Native gain is not established.", + "title": "Repeated confirmation v35 mean 993447588.8" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "7adc885082f3e0c0", + "event_kind": "finding", + "evidence_refs": [ + "research/v43-confirmation-summary.json", + "research/wake17-summary.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:ba4322a5a7e436ccdc5092c561419b48e3cde803d50f86789beeca073b6be5fc", + "comparison_scope": "ad-placement:local:v43-confirmation-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/v43-confirmation-summary.json", + "higher_is_better": true, + "input_revision": "sha256:67c3352c9f08cf20fc4aaa04abfaa0dd74d91c34f556f8c8effd5dce2f370b66", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 993488109.2125 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T00:57:10Z", + "result_id": "finding_2806d4806015", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v43 averaged 993488109.2125 over40 new cases with two4.70-second searches per case. ABBA/BAAB order and case-group analysis preserve repeated observations. All80 source runs passed tester, geometry and resources. Native gain is not established.", + "title": "Repeated confirmation v43 mean 993488109.2125" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "f6af54d905b7e190", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T01:05:21Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 tests. Frozen v43 repeated twice on40 new cases: delta40520 vs v35 (23wins). Uncertainty retained; native incumbent v35.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "9490588fa75ce391", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T01:05:21Z", + "result_id": "node_9a8e7c2231a3", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Native fixed50: v35 49596964560 (50/50), v43 49581803916 (50/50). v35 is the measured incumbent; controlled optimization continues.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "ff97a3160251f5fc", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T01:06:25+00:00", + "result_id": "node_9a8e7c2231a3", + "run_id": "explore-result:1369ed34ee7e959b66ddabbe38c084ff60ef8b961b210336223f9311fab109fb", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Native fixed50: v35 49596964560 (50/50), v43 49581803916 (50/50). v35 is the measured incumbent; controlled optimization continues.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "93c6dcf7d04eddbd", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-confirmation-native" + ], + "execution_observation": { + "candidate_sha256": "ba4322a5a7e436ccdc5092c561419b48e3cde803d50f86789beeca073b6be5fc", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "evidence:ad-confirmation-native" + ], + "executed_method_id": "ad-placement.native.fixed50.v1", + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "limit_seconds": null, + "planned_method_id": "ad-placement.native.fixed50.v1", + "result_candidate_sha256": "ba4322a5a7e436ccdc5092c561419b48e3cde803d50f86789beeca073b6be5fc", + "result_ref": "evidence:ad-confirmation-native", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Accepted native evaluation completed; execution resources not exposed." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:ba4322a5a7e436ccdc5092c561419b48e3cde803d50f86789beeca073b6be5fc", + "comparison_scope": "ad-placement:native:current-task:fixed50", + "evaluation_ref": "native-protocol:README-sha256:0051d2a6e88e299efc0c0b0f1a7e8a68b3060c855b55ffce8c6cfa601d4b30b2", + "evidence_ref": "evidence:ad-confirmation-native", + "higher_is_better": true, + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "metric": "total_native_score", + "unit": "points_out_of_50e9", + "value": 49581803916 + } + ], + "node_id": "node_9a8e7c2231a3", + "recorded_at": "2026-10-23T01:06:25+00:00", + "result_id": "finding_1369ed34ee7e959b66ddabbe38c084ff60ef8b961b210336223f9311fab109fb", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50 Applicability: Native fixed50: v35 49596964560 (50/50), v43 49581803916 (50/50). v35 is the measured incumbent; controlled optimization continues. Observation: Frozen v43 scored49581803916 with50/50 passes in submission9a37ce36f4e8. Delta versusv35:-15160644. Retained v35. Interpretation: The actual accepted ID and completed result match the frozen source archive. The trial did not improve the valid incumbent; exactv35 restored. The reviewed repeated protocol was executed on all40 new cases, with two runs per source and case-group analysis; local evidence did not certify native gain. Public case statuses are preserved. Individual scores, execution time, memory and iterations were not exposed; null resources mean unmeasured. P1 remains open. Continue the next scoped probe from the actual measured incumbent.", + "tags": [ + "writeback-result" + ], + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "84fd71c8f0739fe4", + "event_kind": "finding", + "evidence_refs": [ + "research/tournament-training-selection.json", + "research/validation-wake18.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:tournament-train:gxx-O2:4.70s:mirrored-three-way-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/tournament-training-selection.json", + "higher_is_better": true, + "input_revision": "sha256:1f904eafddf84d8d5959dd0f0840733c7333ba453ee283b29934c8791402be18", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992733120.1 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T01:39:50Z", + "result_id": "finding_071d64e0057f", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 averaged 992733120.1 across 20 matched training cases with two 4.70-second searches each. Mirrored three-way order; all 40 scored completed runs passed; an earlier unscored control exit is preserved separately. Winner v46 was frozen before fresh generation. Training selection is separate from fresh and native evidence.", + "title": "Tournament train v35 mean 992733120.1" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "6d11491bec56a556", + "event_kind": "finding", + "evidence_refs": [ + "research/tournament-training-selection.json", + "research/validation-wake18.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:050fef32759af499d8193bf770ac55bdac723222e404e2bb471ad11a793745d0", + "comparison_scope": "ad-placement:local:tournament-train:gxx-O2:4.70s:mirrored-three-way-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/tournament-training-selection.json", + "higher_is_better": true, + "input_revision": "sha256:1f904eafddf84d8d5959dd0f0840733c7333ba453ee283b29934c8791402be18", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992718340.35 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T01:39:52Z", + "result_id": "finding_5e390f0f4c56", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v45 averaged 992718340.35 across 20 matched training cases with two 4.70-second searches each. Mirrored three-way order; all 40 scored completed runs passed; an earlier unscored control exit is preserved separately. Winner v46 was frozen before fresh generation. Training selection is separate from fresh and native evidence.", + "title": "Tournament train v45 mean 992718340.35" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "56f724666b86db6e", + "event_kind": "finding", + "evidence_refs": [ + "research/tournament-training-selection.json", + "research/validation-wake18.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:b10d9186aa79c1bffa3299611902e3ca26a5cbbffd77158439e88cf5feeaa166", + "comparison_scope": "ad-placement:local:tournament-train:gxx-O2:4.70s:mirrored-three-way-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/tournament-training-selection.json", + "higher_is_better": true, + "input_revision": "sha256:1f904eafddf84d8d5959dd0f0840733c7333ba453ee283b29934c8791402be18", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992864801.6 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T01:39:53Z", + "result_id": "finding_be5550a9b416", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v46 averaged 992864801.6 across 20 matched training cases with two 4.70-second searches each. Mirrored three-way order; all 40 scored completed runs passed; an earlier unscored control exit is preserved separately. Winner v46 was frozen before fresh generation. Training selection is separate from fresh and native evidence.", + "title": "Tournament train v46 mean 992864801.6" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "0c587d8530b7565e", + "event_kind": "finding", + "evidence_refs": [ + "research/wake18-summary.json", + "research/validation-wake18.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:tournament-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/wake18-summary.json", + "higher_is_better": true, + "input_revision": "sha256:621b07355ac4d4346de26c1684dead2ed394bb28c8a140f298fd06b39a51a943", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991821295.7375 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T01:39:54Z", + "result_id": "finding_48788bf0a6b4", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "V35 reference mean 991821295.7375 across 40 fresh cases, two 4.70-second searches each. All 80 source runs passed; full ABBA/BAAB comparison and case-group resampling retained. Native feedback is separate.", + "title": "Tournament fresh40 v35 mean 991821295.7375" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "0527aa37bce95185", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T01:39:58Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 tests. v46 repeated twice on 40 fresh cases: delta 268657 vs v35 (23 wins). Native v35 retained; candidate pending native.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "ac9036d92c4ca22d", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T01:41:23+00:00", + "result_id": "node_e388daf41666", + "run_id": "explore-result:2d1b0414d54afa6b1b694ace827c26fbee0214f74d079ffb51d7dc9003adcf17", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 tests. v46 repeated twice on 40 fresh cases: delta 268657 vs v35 (23 wins). Native v35 retained; candidate pending native.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "9eb13b96ddaa904a", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-tournament-fresh40" + ], + "execution_observation": { + "candidate_sha256": "b10d9186aa79c1bffa3299611902e3ca26a5cbbffd77158439e88cf5feeaa166", + "change_reason": null, + "elapsed_seconds": 377.4868699999999, + "evidence_refs": [ + "evidence:ad-tournament-fresh40" + ], + "executed_method_id": "ad-placement.tourney.abba40x2x4p70s.v1", + "exit_code": 0, + "input_revision": "sha256:621b07355ac4d4346de26c1684dead2ed394bb28c8a140f298fd06b39a51a943", + "limit_seconds": 400, + "planned_method_id": "ad-placement.tourney.abba40x2x4p70s.v1", + "result_candidate_sha256": "b10d9186aa79c1bffa3299611902e3ca26a5cbbffd77158439e88cf5feeaa166", + "result_ref": "evidence:ad-tournament-fresh40", + "schema_version": "explore_execution_observation_v0", + "search_iterations": 702666832, + "stop_reason": "All 80 candidate searches completed within 5 seconds each; both repeats on all 40 cases retained." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:b10d9186aa79c1bffa3299611902e3ca26a5cbbffd77158439e88cf5feeaa166", + "comparison_scope": "ad-placement:local:tournament-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "evidence:ad-tournament-fresh40", + "higher_is_better": true, + "input_revision": "sha256:621b07355ac4d4346de26c1684dead2ed394bb28c8a140f298fd06b39a51a943", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992089952.55 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T01:41:23+00:00", + "result_id": "finding_2d1b0414d54afa6b1b694ace827c26fbee0214f74d079ffb51d7dc9003adcf17", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:621b07355ac4d4346de26c1684dead2ed394bb28c8a140f298fd06b39a51a943 Applicability: 4.70-second O2 tests. v46 repeated twice on 40 fresh cases: delta 268657 vs v35 (23 wins). Native v35 retained; candidate pending native. Observation: v46 averaged 992089952.55 over 40 fresh cases with two runs each, delta 268656.8125 versus v35, 23/40 wins. All 80 candidate runs passed; maximum 4.73125 seconds. Native result unmeasured. Interpretation: Winner frozen after 120 mirrored training runs, before fresh generation. All 160 fresh runs and both cohorts are retained. ABBA/BAAB ordering balanced positions. Runs are averaged within each input; resampling keeps both repetitions together across 40 cases. Exploratory 95% interval [11222.4375,538625.3875] omits residual temporal dependence, training selection and native shift. Repeat differences do not identify scheduling causes. Separate diagnostics verified actual selection entries in both epochs. All 293 completed scored runs passed. One earlier control exit remains unexplained; original artifacts were archived. An isolated full rerun and two targeted sanitizer probes passed; their scores were excluded. Only missing training calls were resumed in order. Positive mean preserves this exact source pending native admission. Native v35 remains; no new native result exists.", + "tags": [ + "writeback-result" + ], + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "ab7016c35e55376f", + "event_kind": "finding", + "evidence_refs": [ + "research/sampling-study-summary.json", + "research/validation-sampling-study.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:sampling-train:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/sampling-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:1f904eafddf84d8d5959dd0f0840733c7333ba453ee283b29934c8791402be18", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992614730.025 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T02:14:27Z", + "result_id": "finding_79ca9ad1800e", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 averaged 992614730.025 on 20 repeated matched training inputs at4.70seconds per search. Fixed prototype, ABBA/BAAB order; all source runs passed tester, independent geometry and resources. Fresh repetitions are grouped by input. Native feedback is separate.", + "title": "Sampling train v35 mean 992614730.025" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "4cc3d75da62857c8", + "event_kind": "finding", + "evidence_refs": [ + "research/sampling-study-summary.json", + "research/validation-sampling-study.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:34ef0856024a888e48a6fcdea2513e204177cba66c812c0757cb22b3c727d546", + "comparison_scope": "ad-placement:local:sampling-train:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/sampling-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:1f904eafddf84d8d5959dd0f0840733c7333ba453ee283b29934c8791402be18", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992999795.475 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T02:14:29Z", + "result_id": "finding_888c83d6ce2d", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v47 averaged 992999795.475 on 20 repeated matched training inputs at4.70seconds per search. Fixed prototype, ABBA/BAAB order; all source runs passed tester, independent geometry and resources. Fresh repetitions are grouped by input. Native feedback is separate.", + "title": "Sampling train v47 mean 992999795.475" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "507b1056236b9bb0", + "event_kind": "finding", + "evidence_refs": [ + "research/sampling-study-summary.json", + "research/validation-sampling-study.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:8af4a746c0595ba9b7d152bbd04387828afdcf84ea424c964db0e0f52d47829c", + "comparison_scope": "ad-placement:local:sampling-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/sampling-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:107af1df3f9a868c20802681ee224b2a50e33a824557fcb0134d4b243a2b3575", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992962660.0625 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T02:14:29Z", + "result_id": "finding_676f75a70dc0", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v35 averaged 992962660.0625 on 40 repeated fresh inputs at4.70seconds per search. Fixed prototype, ABBA/BAAB order; all source runs passed tester, independent geometry and resources. Fresh repetitions are grouped by input. Native feedback is separate.", + "title": "Sampling fresh40 v35 mean 992962660.0625" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "d98cfa680025bfe6", + "event_kind": "finding", + "evidence_refs": [ + "research/sampling-study-summary.json", + "research/validation-sampling-study.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:34ef0856024a888e48a6fcdea2513e204177cba66c812c0757cb22b3c727d546", + "comparison_scope": "ad-placement:local:sampling-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/sampling-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:107af1df3f9a868c20802681ee224b2a50e33a824557fcb0134d4b243a2b3575", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992798208 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T02:14:30Z", + "result_id": "finding_d6441dcb44d7", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v47 averaged 992798208 on 40 repeated fresh inputs at4.70seconds per search. Fixed prototype, ABBA/BAAB order; all source runs passed tester, independent geometry and resources. Fresh repetitions are grouped by input. Native feedback is separate.", + "title": "Sampling fresh40 v47 mean 992798208" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "2ab6117efe85375c", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T02:14:36Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 tests. Frozen v47 repeated twice on40 fresh cases: delta -164452 vs v35 (16 wins). Native v46 retained; v47 parked.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "26f4b1a4b28bf662", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T02:14:37Z", + "result_id": "node_9a8e7c2231a3", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Native fixed50: v35 49596964560 (50/50), v46 49610566071 (50/50). v46 is the measured incumbent; controlled optimization continues.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "2069a48b2bb56bfc", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T02:20:06+00:00", + "result_id": "node_9a8e7c2231a3", + "run_id": "explore-result:164195bfca2a2cfd9aaeb26b7a8588a40f3f7938e5658ece5d104f2e398aaf68", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Native fixed50: v35 49596964560 (50/50), v46 49610566071 (50/50). v46 is the measured incumbent; controlled optimization continues.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "2755c8dafced6136", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-tournament-native" + ], + "execution_observation": { + "candidate_sha256": "b10d9186aa79c1bffa3299611902e3ca26a5cbbffd77158439e88cf5feeaa166", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "evidence:ad-tournament-native" + ], + "executed_method_id": "ad-placement.native.fixed50.v1", + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "limit_seconds": null, + "planned_method_id": "ad-placement.native.fixed50.v1", + "result_candidate_sha256": "b10d9186aa79c1bffa3299611902e3ca26a5cbbffd77158439e88cf5feeaa166", + "result_ref": "evidence:ad-tournament-native", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Accepted native evaluation completed; execution resources not exposed." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:b10d9186aa79c1bffa3299611902e3ca26a5cbbffd77158439e88cf5feeaa166", + "comparison_scope": "ad-placement:native:current-task:fixed50", + "evaluation_ref": "native-protocol:README-sha256:0051d2a6e88e299efc0c0b0f1a7e8a68b3060c855b55ffce8c6cfa601d4b30b2", + "evidence_ref": "evidence:ad-tournament-native", + "higher_is_better": true, + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "metric": "total_native_score", + "unit": "points_out_of_50e9", + "value": 49610566071 + } + ], + "node_id": "node_9a8e7c2231a3", + "recorded_at": "2026-10-23T02:20:06+00:00", + "result_id": "finding_164195bfca2a2cfd9aaeb26b7a8588a40f3f7938e5658ece5d104f2e398aaf68", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50 Applicability: Native fixed50: v35 49596964560 (50/50), v46 49610566071 (50/50). v46 is the measured incumbent; controlled optimization continues. Observation: Frozen v46 scored 49610566071 with 50/50 passes in submission c5eccfd1d463. Delta versus v35: 13601511. Retained v46. Interpretation: The actual accepted ID and completed result match the exact v46 source archive. The valid native score improved and exact v46 is retained. V46 was selected from the earlier repeated frequency study, independently of current v47 outcomes. Native staging did not change local participant snapshots or binaries. The current 40-case repeated v47 comparison remains local evidence against frozen v35; it does not certify improvement over a changed native incumbent. Public case statuses are preserved. Individual native scores, execution time, memory and iterations were not exposed; null resources mean unmeasured. Earlier local failure evidence and all regressions remain archived. P1 stays open.", + "tags": [ + "writeback-result" + ], + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "bb699cb7ce87b96f", + "event_kind": "finding", + "evidence_refs": [ + "research/tailtemp-study-summary.json", + "research/validation-tailtemp-study.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:b10d9186aa79c1bffa3299611902e3ca26a5cbbffd77158439e88cf5feeaa166", + "comparison_scope": "ad-placement:local:tailtemp-train:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/tailtemp-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:1f904eafddf84d8d5959dd0f0840733c7333ba453ee283b29934c8791402be18", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 993041099.45 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T02:54:03Z", + "result_id": "finding_85bf7a6735db", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v46 averaged 993041099.45 on 20 repeated matched training inputs at4.70seconds per search. Fixed prototype, ABBA/BAAB order; all source runs passed tester, independent geometry and resources. Fresh repetitions are grouped by input. Native feedback is separate.", + "title": "Tail temperature train v46 mean 993041099.45" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "2117b441254e84e4", + "event_kind": "finding", + "evidence_refs": [ + "research/tailtemp-study-summary.json", + "research/validation-tailtemp-study.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:2dbcc37b69651e68b9f2ef75cdf800e91b75798d8fcc70f228fbb4e4d5318d5b", + "comparison_scope": "ad-placement:local:tailtemp-train:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/tailtemp-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:1f904eafddf84d8d5959dd0f0840733c7333ba453ee283b29934c8791402be18", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992784089.4 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T02:54:05Z", + "result_id": "finding_5b0b3cb0f3d9", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v48 averaged 992784089.4 on 20 repeated matched training inputs at4.70seconds per search. Fixed prototype, ABBA/BAAB order; all source runs passed tester, independent geometry and resources. Fresh repetitions are grouped by input. Native feedback is separate.", + "title": "Tail temperature train v48 mean 992784089.4" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "55dd288103b4042e", + "event_kind": "finding", + "evidence_refs": [ + "research/tailtemp-study-summary.json", + "research/validation-tailtemp-study.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:b10d9186aa79c1bffa3299611902e3ca26a5cbbffd77158439e88cf5feeaa166", + "comparison_scope": "ad-placement:local:tailtemp-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/tailtemp-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:2317f4eae3b497282d2cce078db2a8a0babb8ad8fb230d2c93291d75b038afa1", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992481289.4125 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T02:54:06Z", + "result_id": "finding_79bda87814d8", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v46 averaged 992481289.4125 on 40 repeated fresh inputs at4.70seconds per search. Fixed prototype, ABBA/BAAB order; all source runs passed tester, independent geometry and resources. Fresh repetitions are grouped by input. Native feedback is separate.", + "title": "Tail temperature fresh40 v46 mean 992481289.4125" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "398bf82f45178277", + "event_kind": "finding", + "evidence_refs": [ + "research/tailtemp-study-summary.json", + "research/validation-tailtemp-study.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:2dbcc37b69651e68b9f2ef75cdf800e91b75798d8fcc70f228fbb4e4d5318d5b", + "comparison_scope": "ad-placement:local:tailtemp-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/tailtemp-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:2317f4eae3b497282d2cce078db2a8a0babb8ad8fb230d2c93291d75b038afa1", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992529211.275 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T02:54:07Z", + "result_id": "finding_ee6325d7b531", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v48 averaged 992529211.275 on 40 repeated fresh inputs at4.70seconds per search. Fixed prototype, ABBA/BAAB order; all source runs passed tester, independent geometry and resources. Fresh repetitions are grouped by input. Native feedback is separate.", + "title": "Tail temperature fresh40 v48 mean 992529211.275" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "e8066347bdd50102", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T02:54:08Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 tests. Frozen v48 repeated twice on40 fresh inputs: delta 47922 vs v46 (21 wins). Native v46 retained; v48 pending native.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "c4e9b5849879edc2", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T02:56:05+00:00", + "result_id": "node_e388daf41666", + "run_id": "explore-result:3091d40a1198c4f6dac13e4c4cc808f6f3b57d21a4dc0b67edaf807c8a6a23cb", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 tests. Frozen v48 repeated twice on40 fresh inputs: delta 47922 vs v46 (21 wins). Native v46 retained; v48 pending native.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "c3ae8d9b94b74d1d", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-tailtemp-fresh40" + ], + "execution_observation": { + "candidate_sha256": "2dbcc37b69651e68b9f2ef75cdf800e91b75798d8fcc70f228fbb4e4d5318d5b", + "change_reason": null, + "elapsed_seconds": 377.48513000000014, + "evidence_refs": [ + "evidence:ad-tailtemp-fresh40" + ], + "executed_method_id": "ad-placement.tailtemp.abba40x2x4p70s.v1", + "exit_code": 0, + "input_revision": "sha256:2317f4eae3b497282d2cce078db2a8a0babb8ad8fb230d2c93291d75b038afa1", + "limit_seconds": 400, + "planned_method_id": "ad-placement.tailtemp.abba40x2x4p70s.v1", + "result_candidate_sha256": "2dbcc37b69651e68b9f2ef75cdf800e91b75798d8fcc70f228fbb4e4d5318d5b", + "result_ref": "evidence:ad-tailtemp-fresh40", + "schema_version": "explore_execution_observation_v0", + "search_iterations": 600371280, + "stop_reason": "All80 candidate searches completed within5seconds each; both repetitions on all40 inputs retained." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:2dbcc37b69651e68b9f2ef75cdf800e91b75798d8fcc70f228fbb4e4d5318d5b", + "comparison_scope": "ad-placement:local:tailtemp-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "evidence:ad-tailtemp-fresh40", + "higher_is_better": true, + "input_revision": "sha256:2317f4eae3b497282d2cce078db2a8a0babb8ad8fb230d2c93291d75b038afa1", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992529211.275 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T02:56:05+00:00", + "result_id": "finding_3091d40a1198c4f6dac13e4c4cc808f6f3b57d21a4dc0b67edaf807c8a6a23cb", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:2317f4eae3b497282d2cce078db2a8a0babb8ad8fb230d2c93291d75b038afa1 Applicability: 4.70-second O2 tests. Frozen v48 repeated twice on40 fresh inputs: delta 47922 vs v46 (21 wins). Native v46 retained; v48 pending native. Observation: Frozen v48 averaged 992529211.275 on40 fresh inputs, two runs each; delta 47921.8625 vs v46, 21/40 wins. All80 candidate runs passed; maximum 4.73081 seconds. Interpretation: One prototype fixed before quality scores; both complete cohorts and all160 fresh runs retained. Balanced ABBA/BAAB positions; average repetitions within input, resample both together across40 case groups. Exploratory95% interval [-296899.475,384691.5125] omits residual timing dependence, campaign selection and native shift. Higher endpoint raises temperatures by4^frac throughout each epoch. Separate short diagnostics reached all bins in both epochs; differing trajectories prevent causal acceptance-rate attribution. All249 current scored runs passed tester, geometry and resources with every child status recorded. Geometry safety evidence reused; prior unexplained failure remains archived. Offline cleanup gained under10 points on saved layouts; those240 transformed outputs are a separate scope. Local evidence does not certify a native gain. Original P1 remains open.", + "tags": [ + "writeback-result" + ], + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "340d1b3beb4c8a24", + "event_kind": "finding", + "evidence_refs": [ + "research/curve-study-summary.json", + "research/validation-curve-study.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:b10d9186aa79c1bffa3299611902e3ca26a5cbbffd77158439e88cf5feeaa166", + "comparison_scope": "ad-placement:local:curve-train:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/curve-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:1f904eafddf84d8d5959dd0f0840733c7333ba453ee283b29934c8791402be18", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992851409.25 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T03:28:07Z", + "result_id": "finding_65c7492e3bf8", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v46 averaged 992851409.25 on 20 repeated matched training inputs at4.70seconds per search. Fixed prototype, ABBA/BAAB order; all source runs passed tester, independent geometry and resources. Fresh repetitions are grouped by input. Native feedback is separate.", + "title": "Cooling curve train v46 mean 992851409.25" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "f9e4f9fe3bbdbef2", + "event_kind": "finding", + "evidence_refs": [ + "research/curve-study-summary.json", + "research/validation-curve-study.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:ecc32e1cc02626934f206f9412e0b0d5db847266f68f7a8ba63efb24ee3dec40", + "comparison_scope": "ad-placement:local:curve-train:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/curve-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:1f904eafddf84d8d5959dd0f0840733c7333ba453ee283b29934c8791402be18", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992081110.35 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T03:28:09Z", + "result_id": "finding_d4bf30de9f81", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v49 averaged 992081110.35 on 20 repeated matched training inputs at4.70seconds per search. Fixed prototype, ABBA/BAAB order; all source runs passed tester, independent geometry and resources. Fresh repetitions are grouped by input. Native feedback is separate.", + "title": "Cooling curve train v49 mean 992081110.35" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "3edddfbca937ccbf", + "event_kind": "finding", + "evidence_refs": [ + "research/curve-study-summary.json", + "research/validation-curve-study.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:b10d9186aa79c1bffa3299611902e3ca26a5cbbffd77158439e88cf5feeaa166", + "comparison_scope": "ad-placement:local:curve-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/curve-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:46dc850f6872fb6a12b69247f4434844d5a5b5dda3de1d7ae52e6d39a0fc259b", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992946888.2375 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T03:28:10Z", + "result_id": "finding_33b8bb62e64d", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v46 averaged 992946888.2375 on 40 repeated fresh inputs at4.70seconds per search. Fixed prototype, ABBA/BAAB order; all source runs passed tester, independent geometry and resources. Fresh repetitions are grouped by input. Native feedback is separate.", + "title": "Cooling curve fresh40 v46 mean 992946888.2375" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "2957446187f6f4f0", + "event_kind": "finding", + "evidence_refs": [ + "research/curve-study-summary.json", + "research/validation-curve-study.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:ecc32e1cc02626934f206f9412e0b0d5db847266f68f7a8ba63efb24ee3dec40", + "comparison_scope": "ad-placement:local:curve-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/curve-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:46dc850f6872fb6a12b69247f4434844d5a5b5dda3de1d7ae52e6d39a0fc259b", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992789610.6125 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T03:28:11Z", + "result_id": "finding_b5a8307462a4", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v49 averaged 992789610.6125 on 40 repeated fresh inputs at4.70seconds per search. Fixed prototype, ABBA/BAAB order; all source runs passed tester, independent geometry and resources. Fresh repetitions are grouped by input. Native feedback is separate.", + "title": "Cooling curve fresh40 v49 mean 992789610.6125" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "80742bb212a3b6ea", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T03:28:12Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 tests. Frozen v49 repeated twice on40 fresh inputs: delta -157278 vs v46 (15 wins). Native v46 retained; v49 parked.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "d474e49d04ffacfe", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T03:28:13Z", + "result_id": "node_9a8e7c2231a3", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Native fixed50: v46 49610566071 (50/50), v48 49594102114 (50/50). v46 is the measured incumbent; optimization continues.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "09b4bd18887e1e7d", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T03:29:49+00:00", + "result_id": "node_9a8e7c2231a3", + "run_id": "explore-result:aa6212a0a4761cef40968663b679cc991426385d1aebf5435a6d943ea1e2530b", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Native fixed50: v46 49610566071 (50/50), v48 49594102114 (50/50). v46 is the measured incumbent; optimization continues.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "1f0298259e8d76aa", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-tailtemp-native" + ], + "execution_observation": { + "candidate_sha256": "2dbcc37b69651e68b9f2ef75cdf800e91b75798d8fcc70f228fbb4e4d5318d5b", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "evidence:ad-tailtemp-native" + ], + "executed_method_id": "ad-placement.native.fixed50.v1", + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "limit_seconds": null, + "planned_method_id": "ad-placement.native.fixed50.v1", + "result_candidate_sha256": "2dbcc37b69651e68b9f2ef75cdf800e91b75798d8fcc70f228fbb4e4d5318d5b", + "result_ref": "evidence:ad-tailtemp-native", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Accepted native evaluation completed; execution resources not exposed." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:2dbcc37b69651e68b9f2ef75cdf800e91b75798d8fcc70f228fbb4e4d5318d5b", + "comparison_scope": "ad-placement:native:current-task:fixed50", + "evaluation_ref": "native-protocol:README-sha256:0051d2a6e88e299efc0c0b0f1a7e8a68b3060c855b55ffce8c6cfa601d4b30b2", + "evidence_ref": "evidence:ad-tailtemp-native", + "higher_is_better": true, + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "metric": "total_native_score", + "unit": "points_out_of_50e9", + "value": 49594102114 + } + ], + "node_id": "node_9a8e7c2231a3", + "recorded_at": "2026-10-23T03:29:49+00:00", + "result_id": "finding_aa6212a0a4761cef40968663b679cc991426385d1aebf5435a6d943ea1e2530b", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50 Applicability: Native fixed50: v46 49610566071 (50/50), v48 49594102114 (50/50). v46 is the measured incumbent; optimization continues. Observation: Frozen v48 scored 49594102114 with 50/50 passes in submission 34532de3ac7d. Delta versus v46: -16463957. Retained v46. Interpretation: Actual acceptedID, completed result and50 public case statuses match the exact v48 archive. Native incumbent remains stronger; exact v46 restored and regression preserved. V48 changed the final temperature and had only small uncertain local fresh support; its training mean was lower. Different input scopes, timing and selection uncertainty prevent a causal explanation for the native discrepancy. The new v49 curve source was frozen only after reconciliation and has separate local evidence; it has no native result. Individual native scores, time, memory and iterations were not exposed; null resources mean unmeasured. Original constraints, P1, all regressions and prior unexplained failure remain preserved. Optimization continues.", + "tags": [ + "writeback-result" + ], + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "54be7c679c9bcce0", + "event_kind": "finding", + "evidence_refs": [ + "research/latecurve-study-summary.json", + "research/validation-latecurve-study.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:b10d9186aa79c1bffa3299611902e3ca26a5cbbffd77158439e88cf5feeaa166", + "comparison_scope": "ad-placement:local:latecurve-train:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/latecurve-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:1f904eafddf84d8d5959dd0f0840733c7333ba453ee283b29934c8791402be18", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992710500.725 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T03:56:48Z", + "result_id": "finding_2b69cdfc54f7", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v46 averaged 992710500.725 on 20 repeated matched training inputs at4.70seconds per search. Fixed prototype, ABBA/BAAB order; all source runs passed tester, independent geometry and resources. Fresh repetitions are grouped by input. Native feedback is separate.", + "title": "Later cooling train v46 mean 992710500.725" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "444e91c49c3cb85a", + "event_kind": "finding", + "evidence_refs": [ + "research/latecurve-study-summary.json", + "research/validation-latecurve-study.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:ae0a291bba542e2ea3e96b7311718eff1a8e716f5c1d72bf9e3167e2c626c98a", + "comparison_scope": "ad-placement:local:latecurve-train:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/latecurve-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:1f904eafddf84d8d5959dd0f0840733c7333ba453ee283b29934c8791402be18", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992755147.9 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T03:56:50Z", + "result_id": "finding_5b75b0c2a3ff", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v50 averaged 992755147.9 on 20 repeated matched training inputs at4.70seconds per search. Fixed prototype, ABBA/BAAB order; all source runs passed tester, independent geometry and resources. Fresh repetitions are grouped by input. Native feedback is separate.", + "title": "Later cooling train v50 mean 992755147.9" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "d4fdd8c7be38d31a", + "event_kind": "finding", + "evidence_refs": [ + "research/latecurve-study-summary.json", + "research/validation-latecurve-study.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:b10d9186aa79c1bffa3299611902e3ca26a5cbbffd77158439e88cf5feeaa166", + "comparison_scope": "ad-placement:local:latecurve-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/latecurve-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:1574ed7c8a680a3ab9da9c71a6208d78eb807f40de206bc31453f2301238e305", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 993052629.35 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T03:56:51Z", + "result_id": "finding_dd9ac8901b2d", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v46 averaged 993052629.35 on 40 repeated fresh inputs at4.70seconds per search. Fixed prototype, ABBA/BAAB order; all source runs passed tester, independent geometry and resources. Fresh repetitions are grouped by input. Native feedback is separate.", + "title": "Later cooling fresh40 v46 mean 993052629.35" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "d23781b78b5e8ae7", + "event_kind": "finding", + "evidence_refs": [ + "research/latecurve-study-summary.json", + "research/validation-latecurve-study.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:ae0a291bba542e2ea3e96b7311718eff1a8e716f5c1d72bf9e3167e2c626c98a", + "comparison_scope": "ad-placement:local:latecurve-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/latecurve-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:1574ed7c8a680a3ab9da9c71a6208d78eb807f40de206bc31453f2301238e305", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 993227783.0625 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T03:56:52Z", + "result_id": "finding_fc78d1cd1f5b", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v50 averaged 993227783.0625 on 40 repeated fresh inputs at4.70seconds per search. Fixed prototype, ABBA/BAAB order; all source runs passed tester, independent geometry and resources. Fresh repetitions are grouped by input. Native feedback is separate.", + "title": "Later cooling fresh40 v50 mean 993227783.0625" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "2bf89abef822aa07", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T03:56:53Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 tests. Frozen v50 repeated twice on40 fresh inputs: delta 175154 vs v46 (25 wins). Native v46 retained; v50 pending native.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "ef822ce917bd8f5e", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T03:58:11+00:00", + "result_id": "node_e388daf41666", + "run_id": "explore-result:666348469ad268c53b3cb2d8cf6400c3d88f6dcdd8088cd5a078ef4095223d11", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 tests. Frozen v50 repeated twice on40 fresh inputs: delta 175154 vs v46 (25 wins). Native v46 retained; v50 pending native.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "e8f5c7fde2f1f776", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-latecurve-fresh40" + ], + "execution_observation": { + "candidate_sha256": "ae0a291bba542e2ea3e96b7311718eff1a8e716f5c1d72bf9e3167e2c626c98a", + "change_reason": null, + "elapsed_seconds": 377.5908299999999, + "evidence_refs": [ + "evidence:ad-latecurve-fresh40" + ], + "executed_method_id": "ad-placement.latecurve.abba40x2x4p70s.v1", + "exit_code": 0, + "input_revision": "sha256:1574ed7c8a680a3ab9da9c71a6208d78eb807f40de206bc31453f2301238e305", + "limit_seconds": 400, + "planned_method_id": "ad-placement.latecurve.abba40x2x4p70s.v1", + "result_candidate_sha256": "ae0a291bba542e2ea3e96b7311718eff1a8e716f5c1d72bf9e3167e2c626c98a", + "result_ref": "evidence:ad-latecurve-fresh40", + "schema_version": "explore_execution_observation_v0", + "search_iterations": 620776528, + "stop_reason": "All80 candidate searches completed within5seconds each; both repetitions on all40 inputs retained." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:ae0a291bba542e2ea3e96b7311718eff1a8e716f5c1d72bf9e3167e2c626c98a", + "comparison_scope": "ad-placement:local:latecurve-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "evidence:ad-latecurve-fresh40", + "higher_is_better": true, + "input_revision": "sha256:1574ed7c8a680a3ab9da9c71a6208d78eb807f40de206bc31453f2301238e305", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 993227783.0625 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T03:58:11+00:00", + "result_id": "finding_666348469ad268c53b3cb2d8cf6400c3d88f6dcdd8088cd5a078ef4095223d11", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:1574ed7c8a680a3ab9da9c71a6208d78eb807f40de206bc31453f2301238e305 Applicability: 4.70-second O2 tests. Frozen v50 repeated twice on40 fresh inputs: delta 175154 vs v46 (25 wins). Native v46 retained; v50 pending native. Observation: Frozen v50 averaged 993227783.0625 on40 fresh inputs, two runs each; delta 175153.7125 vs v46, 25/40 wins. All80 candidate runs passed; maximum 4.73279 seconds. Interpretation: One prototype fixed before quality scores; both complete cohorts and all160 fresh runs retained. Balanced ABBA/BAAB positions; average repetitions within input, resample both together across40 case groups. Exploratory95% interval [-192268.275,545679.8625] omits residual timing dependence, campaign selection and native shift. Only the temperature time exponent changes to frac^1.25; both endpoints stay fixed. Separate short diagnostics reached all bins in both epochs; differing trajectories prevent causal acceptance-rate attribution. All249 current scored runs passed tester, geometry and resources with every child status recorded. Geometry safety evidence reused; prior unexplained failure remains archived. Historical endpoint and earlier-curve regressions remain separate evidence. Local evidence does not certify a native gain. Original P1 remains open.", + "tags": [ + "writeback-result" + ], + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "892f17154f2650d0", + "event_kind": "node", + "evidence_refs": [ + "research/wake24-native-reconciliation.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T04:29:52Z", + "result_id": "node_9a8e7c2231a3", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Native fixed50: v46 49610566071 (50/50), v50 49596571196 (50/50). v46 is the measured incumbent; optimization continues.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "4452bd96b565fa77", + "event_kind": "finding", + "evidence_refs": [ + "research/contact51-study-summary.json", + "research/validation-wake24.json", + "research/execution-v46-contact51-fresh40.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:b10d9186aa79c1bffa3299611902e3ca26a5cbbffd77158439e88cf5feeaa166", + "comparison_scope": "ad-placement:local:contact51-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/contact51-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:b213830bacbd37758a87124a1c39fc755ceaf2d04f76a32ef010cf9a3755c190", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991912080.425 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T04:42:09Z", + "result_id": "finding_c64054a7f352", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v46 averaged 991912080.425 on40 fresh inputs with two4.70-second runs each. Single source fixed before quality; balanced ABBA/BAAB positions and all160 runs retained. All tester, geometry and resource checks passed. Repetitions are grouped by input. Exact execution measurements are attached as an evidence artifact. Native feedback is a separate scope.", + "title": "Filtered contact fresh40 v46 mean 991912080.425" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "59ff88e06b6d36d1", + "event_kind": "finding", + "evidence_refs": [ + "research/contact51-study-summary.json", + "research/validation-wake24.json", + "research/execution-v51-contact51-fresh40.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:193cb26a50832287e84d2c9925d62eb6e6e65e79e8578cb5cd2e25b9ade3400c", + "comparison_scope": "ad-placement:local:contact51-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/contact51-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:b213830bacbd37758a87124a1c39fc755ceaf2d04f76a32ef010cf9a3755c190", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 991935306.8375 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T04:42:10Z", + "result_id": "finding_0d838e5d6930", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v51 averaged 991935306.8375 on40 fresh inputs with two4.70-second runs each. Single source fixed before quality; balanced ABBA/BAAB positions and all160 runs retained. All tester, geometry and resource checks passed. Repetitions are grouped by input. Exact execution measurements are attached as an evidence artifact. Native feedback is a separate scope.", + "title": "Filtered contact fresh40 v51 mean 991935306.8375" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "a32d4d5083ac4106", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T04:42:11Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 tests. Frozen v51 repeated twice on40 fresh inputs: delta 23226 vs v46 (26 wins). Native v46 retained; v51 pending native.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "cdf3028b847861bd", + "event_kind": "node", + "evidence_refs": [ + "research/wake24-native-reconciliation.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T04:45:07+00:00", + "result_id": "node_9a8e7c2231a3", + "run_id": "explore-result:48f7af540ac9fa69d1052363a9432fef78b116414820bc002b34e8ed3dabce8d", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Native fixed50: v46 49610566071 (50/50), v50 49596571196 (50/50). v46 is the measured incumbent; optimization continues.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "5021de58312a01dd", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-latecurve-native" + ], + "execution_observation": { + "candidate_sha256": "ae0a291bba542e2ea3e96b7311718eff1a8e716f5c1d72bf9e3167e2c626c98a", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "evidence:ad-latecurve-native" + ], + "executed_method_id": "ad-placement.native.fixed50.v1", + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "limit_seconds": null, + "planned_method_id": "ad-placement.native.fixed50.v1", + "result_candidate_sha256": "ae0a291bba542e2ea3e96b7311718eff1a8e716f5c1d72bf9e3167e2c626c98a", + "result_ref": "evidence:ad-latecurve-native", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Accepted native evaluation completed; execution resources not exposed." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:ae0a291bba542e2ea3e96b7311718eff1a8e716f5c1d72bf9e3167e2c626c98a", + "comparison_scope": "ad-placement:native:current-task:fixed50", + "evaluation_ref": "native-protocol:README-sha256:0051d2a6e88e299efc0c0b0f1a7e8a68b3060c855b55ffce8c6cfa601d4b30b2", + "evidence_ref": "evidence:ad-latecurve-native", + "higher_is_better": true, + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "metric": "total_native_score", + "unit": "points_out_of_50e9", + "value": 49596571196 + } + ], + "node_id": "node_9a8e7c2231a3", + "recorded_at": "2026-10-23T04:45:07+00:00", + "result_id": "finding_48f7af540ac9fa69d1052363a9432fef78b116414820bc002b34e8ed3dabce8d", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50 Applicability: Native fixed50: v46 49610566071 (50/50), v50 49596571196 (50/50). v46 is the measured incumbent; optimization continues. Observation: Frozen v50 scored 49596571196 with50/50 passes in submission 1fe283023cba. Delta versus v46: -13994875. Exact v46 restored. Interpretation: Actual acceptedID, exact submitted archive, completed score/pass receipt and50 public case statuses agree. V50 later cooling had uncertain repeated local support; it did not improve native score. The measured native incumbent remains v46. Individual native scores, execution time, memory and iterations were not exposed; null resources mean unmeasured. New v51 contact selection has separate preflight and repeated local evidence; it has no native result. Local and native scopes are distinct and do not identify a causal explanation for their discrepancy. Earlier v11/v29/v30 evidence, regressions and the unexplained old failure remain preserved. Original constraints and open P1 continue.", + "tags": [ + "writeback-result" + ], + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "ce5dd9e995ea59bf", + "event_kind": "finding", + "evidence_refs": [ + "research/contact52-study-summary.json", + "research/validation-wake25.json", + "research/execution-v46-contact52-fresh40.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:b10d9186aa79c1bffa3299611902e3ca26a5cbbffd77158439e88cf5feeaa166", + "comparison_scope": "ad-placement:local:contact52-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/contact52-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:ff2bc18d886fb2a7e99e7b1c326d6b0d6e426ef30615351732f5e7563080b218", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992186819.275 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T05:31:00Z", + "result_id": "finding_a45cee036aa8", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v46 averaged 992186819.275 on40 inputs with two4.70-second runs each. All complete cohorts/repetitions retained, balanced ABBA/BAAB order. Official tester, strict geometry, resources and source bindings passed. Execution observations are attached as evidence. This weighted scope is separate from earlier studies and native feedback.", + "title": "weighted v46 mean 992186819.275" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "7642eb0501582b34", + "event_kind": "finding", + "evidence_refs": [ + "research/contact52-study-summary.json", + "research/validation-wake25.json", + "research/execution-v52-contact52-fresh40.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:a2285969fd5b6b23fb66c34dec22dff72d426cddc453c1f7dfb6433be1dc4595", + "comparison_scope": "ad-placement:local:contact52-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/contact52-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:ff2bc18d886fb2a7e99e7b1c326d6b0d6e426ef30615351732f5e7563080b218", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992551797 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T05:31:01Z", + "result_id": "finding_fcb1fc3cda0e", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v52 averaged 992551797 on40 inputs with two4.70-second runs each. All complete cohorts/repetitions retained, balanced ABBA/BAAB order. Official tester, strict geometry, resources and source bindings passed. Execution observations are attached as evidence. This weighted scope is separate from earlier studies and native feedback.", + "title": "weighted v52 mean 992551797" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "19f90f9a6b4ed7d3", + "event_kind": "finding", + "evidence_refs": [ + "research/head20-study-summary.json", + "research/validation-wake25.json", + "research/execution-v51-head20-fresh20.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:193cb26a50832287e84d2c9925d62eb6e6e65e79e8578cb5cd2e25b9ade3400c", + "comparison_scope": "ad-placement:local:head20-fresh20:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/head20-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:dc46d02474913d47f00869235fc30740860287dfa98fca77fcfa51ed3323b829", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 993027529.175 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T05:31:02Z", + "result_id": "finding_c13f5c821f2a", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v51 averaged 993027529.175 on20 inputs with two4.70-second runs each. All complete cohorts/repetitions retained, balanced ABBA/BAAB order. Official tester, strict geometry, resources and source bindings passed. Execution observations are attached as evidence. This head20 scope is separate from earlier studies and native feedback.", + "title": "head20 v51 mean 993027529.175" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "95a6e8326147b8a2", + "event_kind": "finding", + "evidence_refs": [ + "research/head20-study-summary.json", + "research/validation-wake25.json", + "research/execution-v52-head20-fresh20.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:a2285969fd5b6b23fb66c34dec22dff72d426cddc453c1f7dfb6433be1dc4595", + "comparison_scope": "ad-placement:local:head20-fresh20:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/head20-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:dc46d02474913d47f00869235fc30740860287dfa98fca77fcfa51ed3323b829", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992596452.575 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T05:31:03Z", + "result_id": "finding_b59cdb645df9", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v52 averaged 992596452.575 on20 inputs with two4.70-second runs each. All complete cohorts/repetitions retained, balanced ABBA/BAAB order. Official tester, strict geometry, resources and source bindings passed. Execution observations are attached as evidence. This head20 scope is separate from earlier studies and native feedback.", + "title": "head20 v52 mean 992596452.575" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "015476712c5092be", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T05:31:04Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 head20, two runs/source/input: v51 993027529.175, v52 992596452.575. v51 selected by higher mean for native priority; local proxy.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "e748529ee4c76c1b", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T05:31:05Z", + "result_id": "node_9a8e7c2231a3", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Native fixed50: prior v46 49610566071, v51 49607711180 (50/50). v46 is the measured incumbent; optimization continues.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "54bd9171010b6bf0", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T05:31:54+00:00", + "result_id": "node_9a8e7c2231a3", + "run_id": "explore-result:e237aa6930d75b7a073a934ccaed2c21e34fe48551200d3d28e6e0649a9a56e4", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Native fixed50: prior v46 49610566071, v51 49607711180 (50/50). v46 is the measured incumbent; optimization continues.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "efc78212ba39cfec", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-contact-head20-native" + ], + "execution_observation": { + "candidate_sha256": "193cb26a50832287e84d2c9925d62eb6e6e65e79e8578cb5cd2e25b9ade3400c", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "evidence:ad-contact-head20-native" + ], + "executed_method_id": "ad-placement.native.fixed50.v1", + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "limit_seconds": null, + "planned_method_id": "ad-placement.native.fixed50.v1", + "result_candidate_sha256": "193cb26a50832287e84d2c9925d62eb6e6e65e79e8578cb5cd2e25b9ade3400c", + "result_ref": "evidence:ad-contact-head20-native", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Accepted native evaluation completed; execution resources not exposed." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:193cb26a50832287e84d2c9925d62eb6e6e65e79e8578cb5cd2e25b9ade3400c", + "comparison_scope": "ad-placement:native:current-task:fixed50", + "evaluation_ref": "native-protocol:README-sha256:0051d2a6e88e299efc0c0b0f1a7e8a68b3060c855b55ffce8c6cfa601d4b30b2", + "evidence_ref": "evidence:ad-contact-head20-native", + "higher_is_better": true, + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "metric": "total_native_score", + "unit": "points_out_of_50e9", + "value": 49607711180 + } + ], + "node_id": "node_9a8e7c2231a3", + "recorded_at": "2026-10-23T05:31:54+00:00", + "result_id": "finding_e237aa6930d75b7a073a934ccaed2c21e34fe48551200d3d28e6e0649a9a56e4", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50 Applicability: Native fixed50: prior v46 49610566071, v51 49607711180 (50/50). v46 is the measured incumbent; optimization continues. Observation: Frozen v51 scored 49607711180 with 50/50 passes in submission d4c24dbfb780. Delta versus its native incumbent: -2854891. Retained v46. Interpretation: AcceptedID, exact archive, completed score/pass receipt and50 public case statuses agree. A separately frozen common20 study selected this source before native feedback. The earlier weighted and uniform studies and their original decisions remain intact; new common data confirmed the existing v51-first order. Native retention uses a valid strict score improvement over the measured incumbent. Individual case scores, execution time, memory and iterations were not exposed; null resources mean unmeasured. Local comparisons and native fixed50 are separate scopes; positive local evidence does not certify native gain or identify a causal explanation for a discrepancy. All historical regressions, original constraints and prior unexplained failure remain preserved. P1 and optimization continue.", + "tags": [ + "writeback-result" + ], + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "9452cca0a15b4866", + "event_kind": "finding", + "evidence_refs": [ + "research/frequency53-study-summary.json", + "research/validation-wake26.json", + "research/execution-v46-frequency53-fresh40.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:b10d9186aa79c1bffa3299611902e3ca26a5cbbffd77158439e88cf5feeaa166", + "comparison_scope": "ad-placement:local:frequency53-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/frequency53-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:79395eef806d42dc62d90a32cbfd04ce424c172d3fe62c782ae2efabc9b06bb8", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992895961.075 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T06:01:23Z", + "result_id": "finding_bf0837aa3e5d", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v46 averaged 992895961.075 on40 inputs with two4.70-second runs each. All complete cohorts/repetitions retained, balanced ABBA/BAAB order. Official tester, strict geometry, resources and source bindings passed. Execution observations are attached as evidence. This frequency scope is separate from earlier studies and native feedback.", + "title": "frequency v46 mean 992895961.075" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "375d964c554b7091", + "event_kind": "finding", + "evidence_refs": [ + "research/frequency53-study-summary.json", + "research/validation-wake26.json", + "research/execution-v53-frequency53-fresh40.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:37c3ab596d108f54a08697f32cee1aa05f445e7782b4dcbf1a211376f2f6305b", + "comparison_scope": "ad-placement:local:frequency53-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/frequency53-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:79395eef806d42dc62d90a32cbfd04ce424c172d3fe62c782ae2efabc9b06bb8", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992999084.9625 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T06:01:25Z", + "result_id": "finding_cd8f8deefd8c", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v53 averaged 992999084.9625 on40 inputs with two4.70-second runs each. All complete cohorts/repetitions retained, balanced ABBA/BAAB order. Official tester, strict geometry, resources and source bindings passed. Execution observations are attached as evidence. This frequency scope is separate from earlier studies and native feedback.", + "title": "frequency v53 mean 992999084.9625" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "a67f0de8a7c56c40", + "event_kind": "finding", + "evidence_refs": [ + "research/head26-study-summary.json", + "research/validation-wake26.json", + "research/execution-v53-head26-fresh20.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:37c3ab596d108f54a08697f32cee1aa05f445e7782b4dcbf1a211376f2f6305b", + "comparison_scope": "ad-placement:local:head26-fresh20:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/head26-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:4bd4fd6eefba45d06458c7a5dadd170e3699a3509002aa8a56c951b399875257", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992913796.3 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T06:01:26Z", + "result_id": "finding_b2939a05e74a", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v53 averaged 992913796.3 on20 inputs with two4.70-second runs each. All complete cohorts/repetitions retained, balanced ABBA/BAAB order. Official tester, strict geometry, resources and source bindings passed. Execution observations are attached as evidence. This head26 scope is separate from earlier studies and native feedback.", + "title": "head26 v53 mean 992913796.3" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "efa41e5a7beb85a1", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T06:01:27Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 common20: v52 993404713.375, v53 992913796.3, two runs/source/input. v52 chosen for native priority; local proxy.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "49fa9fdd9f5df4e5", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T06:02:43+00:00", + "result_id": "node_e388daf41666", + "run_id": "explore-result:7a568dfe837a8f98e4f5cebdd7ccd00ee5ed11c79f40faa41815154595d6fe6d", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 common20: v52 993404713.375, v53 992913796.3, two runs/source/input. v52 chosen for native priority; local proxy.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "d5b57f2283ea85a6", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-contact-head26" + ], + "execution_observation": { + "candidate_sha256": "a2285969fd5b6b23fb66c34dec22dff72d426cddc453c1f7dfb6433be1dc4595", + "change_reason": null, + "elapsed_seconds": 188.84437000000005, + "evidence_refs": [ + "evidence:ad-contact-head26" + ], + "executed_method_id": "ad-placement.contact-head26.abba20x2x4p70s.v1", + "exit_code": 0, + "input_revision": "sha256:4bd4fd6eefba45d06458c7a5dadd170e3699a3509002aa8a56c951b399875257", + "limit_seconds": 200, + "planned_method_id": "ad-placement.contact-head26.abba20x2x4p70s.v1", + "result_candidate_sha256": "a2285969fd5b6b23fb66c34dec22dff72d426cddc453c1f7dfb6433be1dc4595", + "result_ref": "evidence:ad-contact-head26", + "schema_version": "explore_execution_observation_v0", + "search_iterations": 317745704, + "stop_reason": "All40 source runs completed within5seconds each; both repetitions on all20 inputs retained." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:a2285969fd5b6b23fb66c34dec22dff72d426cddc453c1f7dfb6433be1dc4595", + "comparison_scope": "ad-placement:local:head26-fresh20:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "evidence:ad-contact-head26", + "higher_is_better": true, + "input_revision": "sha256:4bd4fd6eefba45d06458c7a5dadd170e3699a3509002aa8a56c951b399875257", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 993404713.375 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T06:02:43+00:00", + "result_id": "finding_7a568dfe837a8f98e4f5cebdd7ccd00ee5ed11c79f40faa41815154595d6fe6d", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:4bd4fd6eefba45d06458c7a5dadd170e3699a3509002aa8a56c951b399875257 Applicability: 4.70-second O2 common20: v52 993404713.375, v53 992913796.3, two runs/source/input. v52 chosen for native priority; local proxy. Observation: Frozen common20 comparison: v52 mean 993404713.375, v53 mean 992913796.3. Delta v53-minus-v52 -490917.075; v52 selected for native priority. All80 runs passed. Interpretation: This separate prospective follow-up used two immutable validated sources on20 new inputs with two runs per source. The higher complete mean selected native priority; ties retained v52. Both10-input cohorts and all80 runs remain included, balanced ABBA/BAAB. Repetitions stay together when resampling20 case groups. Exploratory95% interval [-912062.275,-113359.375] omits residual timing dependence, earlier adaptive/campaign selection and native shift. Both earlier studies and their original acceptance decisions remain preserved. The new common data can revise an agent-created queue without rewriting prior evidence. It is a local priority decision, not proof of native gain. All249 full and28 short solver runs passed tester, geometry and resources, with all277 child outcomes retained. Original constraints, P1, negative evidence and prior unexplained failure remain preserved.", + "tags": [ + "writeback-result" + ], + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "6bb77f73543d01fb", + "event_kind": "finding", + "evidence_refs": [ + "research/groupsize54-study-summary.json", + "research/execution-v46-groupsize54-fresh40.json", + "research/validation-wake27.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:b10d9186aa79c1bffa3299611902e3ca26a5cbbffd77158439e88cf5feeaa166", + "comparison_scope": "ad-placement:local:groupsize54-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/groupsize54-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:da0cbf3de0fe07653441273f3308a72422d55eb8274f26e43e4a4dc5f53f56ee", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992349289.6875 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T06:36:28Z", + "result_id": "finding_a962219b7fb5", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v46 averaged 992349289.6875 on40 fresh inputs with two4.70-second runs each. Complete cohorts/repetitions retained;balanced ABBA/BAAB. Tester,strict geometry,resources and source bindings passed;execution observation attached as evidence. This local scope is distinct from native evaluation.", + "title": "Group size fresh40 v46 mean 992349289.6875" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "1e8e1af5464c6ccc", + "event_kind": "finding", + "evidence_refs": [ + "research/groupsize54-study-summary.json", + "research/execution-v54-groupsize54-fresh40.json", + "research/validation-wake27.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:7e58fb44dc48f0c6dc2420dd835699b4505fa6cdfceaef180b017b0013fd9aa0", + "comparison_scope": "ad-placement:local:groupsize54-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/groupsize54-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:da0cbf3de0fe07653441273f3308a72422d55eb8274f26e43e4a4dc5f53f56ee", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992533387.6625 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T06:36:31Z", + "result_id": "finding_fd45443197f0", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v54 averaged 992533387.6625 on40 fresh inputs with two4.70-second runs each. Complete cohorts/repetitions retained;balanced ABBA/BAAB. Tester,strict geometry,resources and source bindings passed;execution observation attached as evidence. This local scope is distinct from native evaluation.", + "title": "Group size fresh40 v54 mean 992533387.6625" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "b5e07c473bb14839", + "event_kind": "finding", + "evidence_refs": [ + "research/alt27-study-summary.json", + "research/execution-v53-alt27-fresh20.json", + "research/validation-wake27.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:37c3ab596d108f54a08697f32cee1aa05f445e7782b4dcbf1a211376f2f6305b", + "comparison_scope": "ad-placement:local:alt27-fresh20:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/alt27-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:e20a47eeaf052884ca70ac669a6d087786c61ea2b710c6aeba24bd223ca2c1be", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 993284403.275 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T06:36:32Z", + "result_id": "finding_1fde58f6a7e5", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v53 averaged 993284403.275 on20 new common inputs,two4.70-second runs per input. All80 runs and both cohorts retained,balanced order;tester,geometry,resources and source bindings passed. Ranks immutable alternatives,not a causal count effect or native gain.", + "title": "Alternative common20 v53 mean 993284403.275" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "0a66f5b760e7c0e9", + "event_kind": "finding", + "evidence_refs": [ + "research/alt27-study-summary.json", + "research/execution-v54-alt27-fresh20.json", + "research/validation-wake27.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:7e58fb44dc48f0c6dc2420dd835699b4505fa6cdfceaef180b017b0013fd9aa0", + "comparison_scope": "ad-placement:local:alt27-fresh20:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/alt27-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:e20a47eeaf052884ca70ac669a6d087786c61ea2b710c6aeba24bd223ca2c1be", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 993463934.15 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T06:36:33Z", + "result_id": "finding_0393d0e74ba7", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v54 averaged 993463934.15 on20 new common inputs,two4.70-second runs per input. All80 runs and both cohorts retained,balanced order;tester,geometry,resources and source bindings passed. Ranks immutable alternatives,not a causal count effect or native gain.", + "title": "Alternative common20 v54 mean 993463934.15" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "95cf97122fbb60b8", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T06:36:35Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "4.70-second O2 common20: v53 993284403.275,v54 993463934.15,two runs/source/input. v54 preferred for future native priority.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "9b28d7e010243c84", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T06:36:36Z", + "result_id": "node_9a8e7c2231a3", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Native fixed50: priorv46 49610566071, v52 49615495839 (50/50). v52 is the measured incumbent; optimization continues.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "4122636ff15db1c3", + "event_kind": "node", + "evidence_refs": [ + "research/EXPERIMENTS.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T06:38:19+00:00", + "result_id": "node_9a8e7c2231a3", + "run_id": "explore-result:467fe5c4d1ae1f80f4c3fb544a9bc7018a6f8c21336a48bea71d485dac6dbe26", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Native fixed50: priorv46 49610566071, v52 49615495839 (50/50). v52 is the measured incumbent; optimization continues.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "29e1c29fb49af34f", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-contact52-native" + ], + "execution_observation": { + "candidate_sha256": "a2285969fd5b6b23fb66c34dec22dff72d426cddc453c1f7dfb6433be1dc4595", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "evidence:ad-contact52-native" + ], + "executed_method_id": "ad-placement.native.fixed50.v1", + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "limit_seconds": null, + "planned_method_id": "ad-placement.native.fixed50.v1", + "result_candidate_sha256": "a2285969fd5b6b23fb66c34dec22dff72d426cddc453c1f7dfb6433be1dc4595", + "result_ref": "evidence:ad-contact52-native", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Accepted native evaluation completed;execution resources not exposed." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:a2285969fd5b6b23fb66c34dec22dff72d426cddc453c1f7dfb6433be1dc4595", + "comparison_scope": "ad-placement:native:current-task:fixed50", + "evaluation_ref": "native-protocol:README-sha256:0051d2a6e88e299efc0c0b0f1a7e8a68b3060c855b55ffce8c6cfa601d4b30b2", + "evidence_ref": "evidence:ad-contact52-native", + "higher_is_better": true, + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "metric": "total_native_score", + "unit": "points_out_of_50e9", + "value": 49615495839 + } + ], + "node_id": "node_9a8e7c2231a3", + "recorded_at": "2026-10-23T06:38:19+00:00", + "result_id": "finding_467fe5c4d1ae1f80f4c3fb544a9bc7018a6f8c21336a48bea71d485dac6dbe26", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50 Applicability: Native fixed50: priorv46 49610566071, v52 49615495839 (50/50). v52 is the measured incumbent; optimization continues. Observation: Frozenv52 scored 49615495839 with 50/50 passes in submission 810f0a503b12. Delta vs prior incumbent 4929768. Retained v52. Interpretation: Actual acceptedID,exact submitted archive,completed score/pass receipt and50 public case statuses agree. The prior common26 comparison selectedv52 before native feedback. This result belongs to immutablev52,not the new localv54 study. Adoption requires a valid strict native gain. Native individual scores,execution time,memory and iterations were not exposed;null resources mean unmeasured. V54 has separate repeated local evidence against frozenv46 and no native result. Different input scopes,timing and prior selection prevent a causal explanation for local/native discrepancies. All original constraints,historical regressions and prior unexplained failure remain preserved. P1 and optimization continue.", + "tags": [ + "writeback-result" + ], + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "1718ba774fd3be22", + "event_kind": "finding", + "evidence_refs": [ + "research/groupsize55-study-summary.json", + "research/execution-v52-groupsize55-fresh40.json", + "research/validation-wake28.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:a2285969fd5b6b23fb66c34dec22dff72d426cddc453c1f7dfb6433be1dc4595", + "comparison_scope": "ad-placement:local:groupsize55-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/groupsize55-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:83f63d612c1af2f1413b727894f704c869030fb43f3386f5c8b5bb5b66803bb0", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992594143.875 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T07:09:45Z", + "result_id": "finding_10bc4bca6040", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v52 averaged 992594143.875 on40 inputs,two4.70-second runs each. Complete cohorts/repetitions retained;balanced ABBA/BAAB. Tester,strict geometry,resources and source bindings passed. Local scope is separate from native fixed50.", + "title": "groupsize55-fresh40 v52 mean 992594143.875" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "84ef4f82a650b21d", + "event_kind": "finding", + "evidence_refs": [ + "research/groupsize55-study-summary.json", + "research/execution-v55-groupsize55-fresh40.json", + "research/validation-wake28.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:e6649c494e450e0e77919b7452d707b0235c8038a76849aed62bace0812842bf", + "comparison_scope": "ad-placement:local:groupsize55-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/groupsize55-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:83f63d612c1af2f1413b727894f704c869030fb43f3386f5c8b5bb5b66803bb0", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992653449.8125 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T07:09:46Z", + "result_id": "finding_38f5d3e77a82", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v55 averaged 992653449.8125 on40 inputs,two4.70-second runs each. Complete cohorts/repetitions retained;balanced ABBA/BAAB. Tester,strict geometry,resources and source bindings passed. Local scope is separate from native fixed50.", + "title": "groupsize55-fresh40 v55 mean 992653449.8125" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "3025cd6579f6931d", + "event_kind": "finding", + "evidence_refs": [ + "research/alt28-study-summary.json", + "research/execution-v54-alt28-fresh20.json", + "research/validation-wake28.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:7e58fb44dc48f0c6dc2420dd835699b4505fa6cdfceaef180b017b0013fd9aa0", + "comparison_scope": "ad-placement:local:alt28-fresh20:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/alt28-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:4947b25d11e0b4daa703aba370bbe52eb7957ba8cb367c646bded10d4355f394", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992104687.2 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T07:09:47Z", + "result_id": "finding_89d85a9d1f77", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v54 averaged 992104687.2 on20 inputs,two4.70-second runs each. Complete cohorts/repetitions retained;balanced ABBA/BAAB. Tester,strict geometry,resources and source bindings passed. Local scope is separate from native fixed50.", + "title": "alt28-fresh20 v54 mean 992104687.2" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "00bd2b0f76df56d5", + "event_kind": "node", + "evidence_refs": [ + "research/wake28-local-report.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T07:09:48Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Common20 source ranking: v54 992104687.2, v55 992204548.875, two4.70-second runs/source/input. Preferred future native source v55;native incumbent remainsv52.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "455cde1f5a67fb72", + "event_kind": "node", + "evidence_refs": [ + "research/wake28-local-report.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T07:11:36+00:00", + "result_id": "node_e388daf41666", + "run_id": "explore-result:29aa819fcdf7ddcd8fd575c34a049f4960b2bfe128ea37f688e856b7bc70af82", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Common20 source ranking: v54 992104687.2, v55 992204548.875, two4.70-second runs/source/input. Preferred future native source v55;native incumbent remainsv52.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "5d713bf72123d06f", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-alt28-common20" + ], + "execution_observation": { + "candidate_sha256": "e6649c494e450e0e77919b7452d707b0235c8038a76849aed62bace0812842bf", + "change_reason": null, + "elapsed_seconds": 188.88683000000003, + "evidence_refs": [ + "evidence:ad-alt28-common20" + ], + "executed_method_id": "ad-placement.alt28.abba20x2x4p70s.v1", + "exit_code": 0, + "input_revision": "sha256:4947b25d11e0b4daa703aba370bbe52eb7957ba8cb367c646bded10d4355f394", + "limit_seconds": 200, + "planned_method_id": "ad-placement.alt28.abba20x2x4p70s.v1", + "result_candidate_sha256": "e6649c494e450e0e77919b7452d707b0235c8038a76849aed62bace0812842bf", + "result_ref": "evidence:ad-alt28-common20", + "schema_version": "explore_execution_observation_v0", + "search_iterations": 325857064, + "stop_reason": "All source runs completed within5seconds each;both repetitions on every input retained." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:e6649c494e450e0e77919b7452d707b0235c8038a76849aed62bace0812842bf", + "comparison_scope": "ad-placement:local:alt28-fresh20:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "evidence:ad-alt28-common20", + "higher_is_better": true, + "input_revision": "sha256:4947b25d11e0b4daa703aba370bbe52eb7957ba8cb367c646bded10d4355f394", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992204548.875 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T07:11:36+00:00", + "result_id": "finding_29aa819fcdf7ddcd8fd575c34a049f4960b2bfe128ea37f688e856b7bc70af82", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:4947b25d11e0b4daa703aba370bbe52eb7957ba8cb367c646bded10d4355f394 Applicability: Common20 source ranking: v54 992104687.2, v55 992204548.875, two4.70-second runs/source/input. Preferred future native source v55;native incumbent remainsv52. Observation: Common20 v54 mean 992104687.2, v55 992204548.875. V55-minus-v54 99861.675, 14/20 wins. Frozen rule selects v55 for native priority. All80 runs passed. Interpretation: Separate prospective comparison of two immutable validated sources on20 new inputs,two runs/source,balanced ABBA/BAAB. V54 is fixed2/nearest fromv46;v55 is fixed3/weighted fromv52. This ranks sources,not a causal size/contact effect. Every cohort and repetition is retained;resample20 whole input groups. Exploratory95% interval [-223579.025,400197.675] omits residual timing dependence,prior campaign selection and native shift. Earlierv55-versusv52 study remains intact: delta59305.9375,24/40 wins,interval spanning zero,repetitions disagree;positive mean retainsv55 as an alternative. Epoch counters verify actual sizes and weighted contacts. Combined249 full and28 short runs passed tester,strict geometry,resources;277 attempts,0failed. Nativev52 remains49615495839,50/50;no new native result. P1,constraints,prior regressions and unexplained failure remain.", + "tags": [ + "writeback-result" + ], + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "bc79bc00c80963d2", + "event_kind": "finding", + "evidence_refs": [ + "research/groupsize56-study-summary.json", + "research/execution-v52-groupsize56-fresh40.json", + "research/validation-wake29.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:a2285969fd5b6b23fb66c34dec22dff72d426cddc453c1f7dfb6433be1dc4595", + "comparison_scope": "ad-placement:local:groupsize56-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/groupsize56-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:d606c1303a7fa38052c67cee5f8d2ff701e2a6e32def5217f53a000fe0a1dc07", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992833883.175 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T07:32:38Z", + "result_id": "finding_fa2332d099c0", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v52 averaged 992833883.175 on40 new inputs,two4.70-second runs per source/input. All cases/repetitions/cohorts retained,balanced order. Tester,geometry,resources and source identity passed. Local source ranking is separate from native fixed50.", + "title": "groupsize56-fresh40 v52 mean 992833883.175" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "f8d81732072c9760", + "event_kind": "finding", + "evidence_refs": [ + "research/groupsize56-study-summary.json", + "research/execution-v56-groupsize56-fresh40.json", + "research/validation-wake29.json" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:fe15b1e79efd2442ae5e5fe6b5cf13c7339d7c80e350802b040c32968470213c", + "comparison_scope": "ad-placement:local:groupsize56-fresh40:gxx-O2:4.70s:paired-two-repeat-v1", + "evaluation_ref": "tester:sha256:2498e6a2d2cfa2435ad541ebe8a284ea576389c53f9a05555b594dcc3a360dcb", + "evidence_ref": "research/groupsize56-study-summary.json", + "higher_is_better": true, + "input_revision": "sha256:d606c1303a7fa38052c67cee5f8d2ff701e2a6e32def5217f53a000fe0a1dc07", + "metric": "mean_satisfaction_score", + "unit": "points_out_of_1e9", + "value": 992767663.1375 + } + ], + "node_id": "node_e388daf41666", + "recorded_at": "2026-10-23T07:32:39Z", + "result_id": "finding_9386464f4ff7", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "v56 averaged 992767663.1375 on40 new inputs,two4.70-second runs per source/input. All cases/repetitions/cohorts retained,balanced order. Tester,geometry,resources and source identity passed. Local source ranking is separate from native fixed50.", + "title": "groupsize56-fresh40 v56 mean 992767663.1375" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "19139df7ed9acaac", + "event_kind": "node", + "evidence_refs": [ + "research/wake29-local-report.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T07:32:40Z", + "result_id": "node_e388daf41666", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Weighted fixed2 fresh40 v52 992833883.175,v56 992767663.1375,two4.70-second runs/source/input;scoped local proxy.", + "title": "Which search variant is best supported by the current fixed-seed 4.7-second local evaluation?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "8ec0070b2b4e23b9", + "event_kind": "node", + "evidence_refs": [ + "research/wake29-local-report.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T07:32:41Z", + "result_id": "node_9a8e7c2231a3", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Native fixed50 v55 49614884448 (50/50);retained v52 49615495839. Optimization continues.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "184ff2316ba0ff83", + "event_kind": "node", + "evidence_refs": [ + "research/wake29-local-report.md" + ], + "goal_id": "ad-explore-test-20261023-18h03", + "node_kind": "question", + "recorded_at": "2026-10-23T07:35:25+00:00", + "result_id": "node_9a8e7c2231a3", + "run_id": "explore-result:4abd75910ea27fd031942df2397cac34b916a7231537eac3eeaf209fa28a4288", + "schema_version": "loopx_explore_result_event_v0", + "status": "exploring", + "summary": "Native fixed50 v55 49614884448 (50/50);retained v52 49615495839. Optimization continues.", + "title": "Which tested candidate has the highest native score on this task?" + }, + { + "agent_id": "ad-explore-test-20261023-18h03-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "eaf4d272f44bbf53", + "event_kind": "finding", + "evidence_refs": [ + "evidence:ad-groupsize55-native" + ], + "execution_observation": { + "candidate_sha256": "e6649c494e450e0e77919b7452d707b0235c8038a76849aed62bace0812842bf", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "evidence:ad-groupsize55-native" + ], + "executed_method_id": "ad-placement.native.fixed50.v1", + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "limit_seconds": null, + "planned_method_id": "ad-placement.native.fixed50.v1", + "result_candidate_sha256": "e6649c494e450e0e77919b7452d707b0235c8038a76849aed62bace0812842bf", + "result_ref": "evidence:ad-groupsize55-native", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Accepted native evaluation completed;execution resources were not exposed." + }, + "goal_id": "ad-explore-test-20261023-18h03", + "measurements": [ + { + "candidate_ref": "sha256:e6649c494e450e0e77919b7452d707b0235c8038a76849aed62bace0812842bf", + "comparison_scope": "ad-placement:native:current-task:fixed50", + "evaluation_ref": "native-protocol:README-sha256:0051d2a6e88e299efc0c0b0f1a7e8a68b3060c855b55ffce8c6cfa601d4b30b2", + "evidence_ref": "evidence:ad-groupsize55-native", + "higher_is_better": true, + "input_revision": "task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50", + "metric": "total_native_score", + "unit": "points_out_of_50e9", + "value": 49614884448 + } + ], + "node_id": "node_9a8e7c2231a3", + "recorded_at": "2026-10-23T07:35:25+00:00", + "result_id": "finding_4abd75910ea27fd031942df2397cac34b916a7231537eac3eeaf209fa28a4288", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: task:e4c58459ec09ddef59890b55ec6e6346fb1aa6cf498444961246b164935d200a:fixed50 Applicability: Native fixed50 v55 49614884448 (50/50);retained v52 49615495839. Optimization continues. Observation: Frozenv55 scored 49614884448,50/50 passes,submission 364f7f9f0c3d. Delta -611391 vs nativev52. Retained v52. Interpretation: AcceptedID,exact immutable archive,completed score/pass receipt and50 unique public case statuses agree. Prior common28 data selectedv55 before native feedback. Adoption requires valid50/50 strict improvement over the measured incumbent. Native individual scores,execution time,memory and iterations were not exposed;null resources mean unmeasured. This delayed result belongs to immutablev55,not the currentv56 files. Isolated localv56 retains frozenv52 as control even if nativev55 is adopted. The complete nonpositivev56 study parks that source. All169 full and28 short local runs passed,all197 solver outcomes retained. Local/native discrepancy does not identify a causal explanation. Original P1,constraints,historical regressions and prior unexplained failure remain;optimization continues.", + "tags": [ + "writeback-result" + ], + "title": "Which tested candidate has the highest native score on this task?" + } + ] +} diff --git a/apps/presentation/dashboard/src/data/fixtures/edgebench-schema-20261025-history.json b/apps/presentation/dashboard/src/data/fixtures/edgebench-schema-20261025-history.json new file mode 100644 index 0000000000..c4b3f81750 --- /dev/null +++ b/apps/presentation/dashboard/src/data/fixtures/edgebench-schema-20261025-history.json @@ -0,0 +1,4767 @@ +{ + "description": "Complete captured Explore display state from a real EdgeBench run. Original wording, event and result identities, recording times, measurements and references are unchanged.", + "source_sha256": "5473191a1f23537a48f38ba82d0dd8ba8ed5ad7daffd944f28910e7d6e19e0ab", + "source_bytes": 271025, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "display_name": "Schemathesis · 2026-10-25", + "provenance": { + "case_id": "schemathesis_config_modernization", + "source_run_id": "schema-7269-test-solmax-r3-20261025-4h05", + "study_id": "pr7269-2x2-solmax-replan3-20261025-04", + "selection": "Record-rich complete collected Explore logs selected after inspecting newer local captures: the 26 October Ad and Schema logs contain only two and four records. This Ad history exercises repeated result versions; this Schema history exercises inherited records across three UTC dates. Selection is for reading complexity, not score. No events are sampled.", + "capture_kind": "terminal_collected_display_state", + "exported_at": "2026-11-04T07:12:13.753444+00:00", + "claim_scope": "diagnostic_only", + "score_countable": false, + "runner_status": "cancelled", + "limitations": [ + "The complete saved Explore log is not the complete solver trajectory or a coverage claim for every attempt.", + "Recorded confirmed labels are source-authored; negative and unknown results remain in their original prose.", + "No explicit parent or relationship edges were recorded. Shared wording, time adjacency and version updates do not establish causal relationships.", + "Historical event times and originating agents are retained, including records inherited by the actual destination run. Recording intervals are not active solver time.", + "Native and local measurement scopes remain distinct. Cancelled or invalid runs do not qualify a scientific comparison." + ], + "redactions": [], + "boundary": "Display-state events only. No authority or credential store, raw session transcript, hidden evaluator, workspace or model weights are included." + }, + "source_counts": { + "event_count": 94, + "event_kinds": { + "node": 47, + "finding": 47 + }, + "statuses": { + "open": 1, + "confirmed": 47, + "resolved": 46 + }, + "first_recorded_at": "2026-10-23T17:50:32+00:00", + "last_recorded_at": "2026-10-25T15:39:04+00:00", + "edge_count": 0, + "parent_records": 0 + }, + "events": [ + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "53b4acf2dfb36b1d", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T17:50:32+00:00", + "result_id": "question_8c9100feb9b1f9f32d20188251396fbe4196e4087fd039d18b689f93df6a914a", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:f8dd7993f74a2a5f", + "schema_version": "loopx_explore_result_event_v0", + "status": "open", + "summary": "TODO.md modernization task, native submission 6a2a5680adaf; no claim about later source versions.", + "title": "Does the integrated candidate satisfy all six roadmap phases in native evaluation?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "e5aa668c0d8f3821", + "event_kind": "finding", + "evidence_refs": [ + "submission:6a2a5680adaf", + "research/api/submission-1/native-result.txt", + "research/api/submission-1/candidate-manifest.json" + ], + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "candidate:18e2e9026fb41992a70ca7e670e4e5d9a6c721b263af2e399f01a9877cb1e5fd", + "comparison_scope": "native:6a2a5680adaf", + "evaluation_ref": "roadmap_structured_json:6a2a5680adaf", + "evidence_ref": "submission:6a2a5680adaf", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "native_score", + "unit": "score", + "value": 0.729591 + } + ], + "node_id": "question_8c9100feb9b1f9f32d20188251396fbe4196e4087fd039d18b689f93df6a914a", + "recorded_at": "2026-10-23T17:50:32+00:00", + "result_id": "finding_f179b89cf341308e0995bd70acbf4396a380072ce8f358826271b0cc80d0d21e", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:cffa87bbcc66afd4", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89 Applicability: TODO.md modernization task, native submission 6a2a5680adaf; no claim about later source versions. Observation: Submission 6a2a5680adaf scored 0.729591; 1/6 phases passed. API namespace passed 26/26 checks. Other phase checks: configuration 14/30, metrics 13/15, transport 14/19, Response 15/18, checks 16/18. Interpretation: The evaluated candidate does not satisfy the full task. The report gives phase totals only. Preserve its source and audit remaining interfaces, prioritizing configuration. Later local fixes are a separate ungraded candidate; wait for eligibility before submitting them.", + "tags": [ + "writeback-result" + ], + "title": "Does the integrated candidate satisfy all six roadmap phases in native evaluation?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "0f01647c4cd7e67e", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T18:02:40+00:00", + "result_id": "question_eed6f130fe08dda4221791ef29f92254b4b8b944bba0a5f02395802283bcf449", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:c0ba6132a387c001", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Local regression suite research/audit/test_audit.py comparing retained candidate 2 with candidate 3; no inference about native per-test failures or overall roadmap completion.", + "title": "Do the audit corrections resolve the reproduced configuration, coverage serializer and response conversion defects?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "545e22ab484b6c7c", + "event_kind": "finding", + "evidence_refs": [ + "research/audit/baseline-provenance.json", + "research/audit/baseline-test-results.txt", + "research/audit/full-test-results.txt", + "research/audit/candidate-3-manifest.json", + "research/audit/test-results.txt" + ], + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:5cb55d8e39c4b0794bd1c70b260a6395b0b3408954da7d9e30ba4037ee5015f5", + "comparison_scope": "local:audit-17:sha256:225665cf18267b6057d6b6ebe3d2b6c88a835d6d3d1e7f46b5a42cfd89f40975", + "evaluation_ref": "pytest:8.4.2:python:3.11:noconftest:audit-only", + "evidence_ref": "research/audit/test-results.txt", + "higher_is_better": true, + "input_revision": "sha256:225665cf18267b6057d6b6ebe3d2b6c88a835d6d3d1e7f46b5a42cfd89f40975", + "metric": "audit_regressions_passed", + "unit": "tests", + "value": 17 + } + ], + "node_id": "question_eed6f130fe08dda4221791ef29f92254b4b8b944bba0a5f02395802283bcf449", + "recorded_at": "2026-10-23T18:02:40+00:00", + "result_id": "finding_93842a0c026556dcd574d85de5af74ad1c63586bc4574c146b41b589b5db744a", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:617f528c41f19ecb", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:225665cf18267b6057d6b6ebe3d2b6c88a835d6d3d1e7f46b5a42cfd89f40975 Applicability: Local regression suite research/audit/test_audit.py comparing retained candidate 2 with candidate 3; no inference about native per-test failures or overall roadmap completion. Observation: Candidate 2 fails 16 of 17 audit regressions; candidate 3 passes all 17, the full 163-test focused suite, and 11 existing regressions. No native submission was made during cooldown. Interpretation: The reproduced local defects are resolved in candidate 3. Validation covers inheritance, helper input errors, selector regexes, serializer coverage selection, WSGI metadata and JSON decoding. This is local evidence, not a native score improvement. Keep the task open and audit the remaining documented interfaces before the next eligible native evaluation.", + "tags": [ + "writeback-result" + ], + "title": "Do the audit corrections resolve the reproduced configuration, coverage serializer and response conversion defects?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "77ec25aa078c8fbd", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T18:12:37+00:00", + "result_id": "question_3e8bf7bbdb778eb97a19bef234f6ac81ef87655068d7345e27277a6fbca437c2", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:0acefa7b90b672a2", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Local 22-test research/execution/test_generation_execution.py suite comparing retained candidate 3 with candidate 4; no inference about native per-test failures.", + "title": "Do configured generation controls reach public loaders, generated tests and runner metrics while preserving explicit overrides?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "fb902148a4581db1", + "event_kind": "finding", + "evidence_refs": [ + "research/execution/baseline-provenance.json", + "research/execution/baseline-test-results.txt", + "research/execution/test-results.txt", + "research/execution/full-test-results.txt", + "research/execution/candidate-4-manifest.json" + ], + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:a56879ccf88d5ee55651a043752481b471b1cf878fa6df5a492f1ea16049def8", + "comparison_scope": "local:execution-22:sha256:e01c9fcea18e08eb568bd850624821d22efaef76f2bd62988c4f231747e6414d", + "evaluation_ref": "pytest:8.4.2:python:3.11:noconftest:execution-only", + "evidence_ref": "research/execution/test-results.txt", + "higher_is_better": true, + "input_revision": "sha256:e01c9fcea18e08eb568bd850624821d22efaef76f2bd62988c4f231747e6414d", + "metric": "execution_regressions_passed", + "unit": "tests", + "value": 22 + } + ], + "node_id": "question_3e8bf7bbdb778eb97a19bef234f6ac81ef87655068d7345e27277a6fbca437c2", + "recorded_at": "2026-10-23T18:12:37+00:00", + "result_id": "finding_7de2eb4bdd0dd2ae592eef15b2d191cad53fd6b922bb978a9a89164f3cdbed48", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:362fd2459995bd70", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:e01c9fcea18e08eb568bd850624821d22efaef76f2bd62988c4f231747e6414d Applicability: Local 22-test research/execution/test_generation_execution.py suite comparing retained candidate 3 with candidate 4; no inference about native per-test failures. Observation: Candidate 3 fails 19 of 22 execution regressions. Candidate 4 passes all 22, the full 185-test focused suite and 7 existing GraphQL regressions. Native history still contains only submission 6a2a5680adaf. Interpretation: The tested configuration propagation gaps are resolved. All 12 loader paths retain config, runtime generation follows configured controls, explicit settings win, and runner metrics use schema configuration. The roadmap remains open because candidate 4 has no native evaluation. Preserve it for the next eligible submission and continue independent interface review.", + "tags": [ + "writeback-result" + ], + "title": "Do configured generation controls reach public loaders, generated tests and runner metrics while preserving explicit overrides?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "7b906a2333954f69", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T18:27:58+00:00", + "result_id": "question_c179b6805a6d2b35cd12d877765348a3314a751a54ab577d064599e0246ba3c2", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:b47eaccf34e85c1b", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Local 14-test research/operations/test_operation_config.py comparison of retained candidate 4 and candidate 5; no inference about hidden native assertions.", + "title": "Does operation resolution preserve current dataclass settings, explicit overrides and programmatic strategies without mutating the source?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "719952fadafdbbb3", + "event_kind": "finding", + "evidence_refs": [ + "research/operations/baseline-provenance.json", + "research/operations/baseline-test-results.txt", + "research/operations/test-results.txt", + "research/operations/full-test-results.txt", + "research/operations/candidate-5-manifest.json" + ], + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:99f46692f5829090eaabd39b77816b9d7b3d16bb01e0f0ee059ffa3684f2febe", + "comparison_scope": "local:operations-14:sha256:855844378181b662fd71a1c99a718b7e88c0193db887e03a65a4ccf2b79b79c2", + "evaluation_ref": "pytest:8.4.2:python:3.11:noconftest:operations-only", + "evidence_ref": "research/operations/test-results.txt", + "higher_is_better": true, + "input_revision": "sha256:855844378181b662fd71a1c99a718b7e88c0193db887e03a65a4ccf2b79b79c2", + "metric": "operation_regressions_passed", + "unit": "tests", + "value": 14 + } + ], + "node_id": "question_c179b6805a6d2b35cd12d877765348a3314a751a54ab577d064599e0246ba3c2", + "recorded_at": "2026-10-23T18:27:58+00:00", + "result_id": "finding_e804545aef66352949d927c8572e0f24d9ad3bef0bafb3c6f4b616b3d717a4b2", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:54a6854518f7abec", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:855844378181b662fd71a1c99a718b7e88c0193db887e03a65a4ccf2b79b79c2 Applicability: Local 14-test research/operations/test_operation_config.py comparison of retained candidate 4 and candidate 5; no inference about hidden native assertions. Observation: Candidate 4 fails 11 of 14 operation regressions; candidate 5 passes all 14, all 199 focused tests and 4 existing utility tests. Native history still contains only submission 6a2a5680adaf. Interpretation: The tested configuration loss is resolved. Current settings survive lookup, ordered overrides and phase inheritance remain effective, custom strategies generate headers, and sources remain isolated. Public factories still validate input. Preserve candidate 5 for the next eligible evaluation; all six targets remain required and no native improvement is claimed.", + "tags": [ + "writeback-result" + ], + "title": "Does operation resolution preserve current dataclass settings, explicit overrides and programmatic strategies without mutating the source?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "955e87f74e68fa88", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T18:49:28+00:00", + "result_id": "question_0d1c42e5adc4be696c3f1ca650fd881c9c14dbe72ad88286f2eba2f5cec2f9b4", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:9b6d65767938143f", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Public native report for submission c1944b9c1f88 matched to candidate-6.zip and its manifest; phase totals only. No hidden assertion attribution or claim of evaluator-version equivalence.", + "title": "Does preserved candidate 6 satisfy all six roadmap phases in native submission c1944b9c1f88?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "44e4ee1d8a0a0dc3", + "event_kind": "finding", + "evidence_refs": [ + "research/serialization/native-result.txt", + "research/serialization/submission-provenance.json", + "research/serialization/candidate-6-manifest.json", + "research/serialization/REPORT.md" + ], + "execution_observation": { + "candidate_sha256": "6fef4bf9809f68810bb51de7e7044083532c58db0c8cfb2cf181af0feea3ebed", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "research/serialization/candidate-6-manifest.json", + "research/serialization/native-result.txt", + "research/serialization/submission-provenance.json" + ], + "executed_method_id": "sforge-submit.native", + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "limit_seconds": null, + "planned_method_id": "sforge-submit.native", + "result_candidate_sha256": "6fef4bf9809f68810bb51de7e7044083532c58db0c8cfb2cf181af0feea3ebed", + "result_ref": "research/serialization/native-result.txt", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Accepted evaluation completed; all six phases remain required." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:6fef4bf9809f68810bb51de7e7044083532c58db0c8cfb2cf181af0feea3ebed", + "comparison_scope": "native:c1944b9c1f88:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/serialization/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "weighted_score", + "unit": "ratio", + "value": 0.786257 + }, + { + "candidate_ref": "sha256:6fef4bf9809f68810bb51de7e7044083532c58db0c8cfb2cf181af0feea3ebed", + "comparison_scope": "native:c1944b9c1f88:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/serialization/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "complete_phases", + "unit": "phases", + "value": 1 + }, + { + "candidate_ref": "sha256:6fef4bf9809f68810bb51de7e7044083532c58db0c8cfb2cf181af0feea3ebed", + "comparison_scope": "native:c1944b9c1f88:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/serialization/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "configuration_checks_passed", + "unit": "checks", + "value": 19 + }, + { + "candidate_ref": "sha256:6fef4bf9809f68810bb51de7e7044083532c58db0c8cfb2cf181af0feea3ebed", + "comparison_scope": "native:c1944b9c1f88:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/serialization/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "api_checks_passed", + "unit": "checks", + "value": 26 + }, + { + "candidate_ref": "sha256:6fef4bf9809f68810bb51de7e7044083532c58db0c8cfb2cf181af0feea3ebed", + "comparison_scope": "native:c1944b9c1f88:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/serialization/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "metrics_checks_passed", + "unit": "checks", + "value": 14 + }, + { + "candidate_ref": "sha256:6fef4bf9809f68810bb51de7e7044083532c58db0c8cfb2cf181af0feea3ebed", + "comparison_scope": "native:c1944b9c1f88:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/serialization/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "transport_checks_passed", + "unit": "checks", + "value": 14 + }, + { + "candidate_ref": "sha256:6fef4bf9809f68810bb51de7e7044083532c58db0c8cfb2cf181af0feea3ebed", + "comparison_scope": "native:c1944b9c1f88:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/serialization/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "response_checks_passed", + "unit": "checks", + "value": 15 + }, + { + "candidate_ref": "sha256:6fef4bf9809f68810bb51de7e7044083532c58db0c8cfb2cf181af0feea3ebed", + "comparison_scope": "native:c1944b9c1f88:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/serialization/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "checks_checks_passed", + "unit": "checks", + "value": 16 + } + ], + "node_id": "question_0d1c42e5adc4be696c3f1ca650fd881c9c14dbe72ad88286f2eba2f5cec2f9b4", + "recorded_at": "2026-10-23T18:49:28+00:00", + "result_id": "finding_e2bbffc1af1a5d9200a0270f3539b762d802b18279fed33e0f19b4da718c5388", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:457f22eb470227b1", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89 Applicability: Public native report for submission c1944b9c1f88 matched to candidate-6.zip and its manifest; phase totals only. No hidden assertion attribution or claim of evaluator-version equivalence. Observation: Candidate 6 scores 0.786257 with 1/6 complete phases: configuration 19/30, API 26/26, metrics 14/15, transport 14/19, response 15/18, checks 16/18. Local validation passes 221 focused tests and 7 existing regressions. Interpretation: The scoped completion question has a negative answer. Retain candidate 6 as the best native-supported source, while keeping the integration Todo and all six requirements open. Candidate 1 scored 0.729591; the new public report gains five configuration checks and one metric check. Reports disclose matching phase totals and weights but no evaluator revision; measurements remain separate rather than claiming a fully controlled causal comparison. Review documented public contracts and reproduce defects locally, prioritizing configuration. Follow server cooldown before the next submission.", + "tags": [ + "writeback-result" + ], + "title": "Does preserved candidate 6 satisfy all six roadmap phases in native submission c1944b9c1f88?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "89e9842a7b2f90ab", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T18:59:03+00:00", + "result_id": "question_46432333668188a71bdbaa119351da1c6a1fb05f7ae7f8bad28c7154b315239d", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:d3f9118e4c6b7329", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "The 33 local configuration boundary cases compared against retained candidate 6 and candidate 7, plus the recorded ten-module integrated suite; no attribution to hidden native assertions.", + "title": "Does configuration loading reject unusable numeric settings and preserve actionable TOML error paths?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "cb77ba44d40b7cc0", + "event_kind": "finding", + "evidence_refs": [ + "research/config-validation/REPORT.md", + "research/config-validation/baseline-provenance.json", + "research/config-validation/full-test-results.txt", + "research/config-validation/candidate-7-manifest.json", + "research/config-validation/evaluation-input-manifest.json" + ], + "execution_observation": { + "candidate_sha256": "536d225cb145205d3ecfc4a8c962153ff6d400c5dd4e4e5f587dff429f2c5e57", + "change_reason": null, + "elapsed_seconds": 4.61, + "evidence_refs": [ + "research/config-validation/candidate-7-manifest.json", + "research/config-validation/evaluation-input-manifest.json", + "research/config-validation/full-test-results.txt" + ], + "executed_method_id": "pytest.focused-roadmap-suite", + "exit_code": 0, + "input_revision": "sha256:5bbb7d9323548b98267932e8a19948e4b98c915aee40a37f47462747ea7c8f5b", + "limit_seconds": null, + "planned_method_id": "pytest.focused-roadmap-suite", + "result_candidate_sha256": "536d225cb145205d3ecfc4a8c962153ff6d400c5dd4e4e5f587dff429f2c5e57", + "result_ref": "research/config-validation/full-test-results.txt", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "All 254 integrated local tests passed. Native evaluation is not yet eligible." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:536d225cb145205d3ecfc4a8c962153ff6d400c5dd4e4e5f587dff429f2c5e57", + "comparison_scope": "local:focused-roadmap-suite:sha256:5bbb7d9323548b98267932e8a19948e4b98c915aee40a37f47462747ea7c8f5b", + "evaluation_ref": "pytest:8.4.2:python:3.11:noconftest:roadmap", + "evidence_ref": "research/config-validation/full-test-results.txt", + "higher_is_better": true, + "input_revision": "sha256:5bbb7d9323548b98267932e8a19948e4b98c915aee40a37f47462747ea7c8f5b", + "metric": "focused_tests_passed", + "unit": "tests", + "value": 254 + } + ], + "node_id": "question_46432333668188a71bdbaa119351da1c6a1fb05f7ae7f8bad28c7154b315239d", + "recorded_at": "2026-10-23T18:59:03+00:00", + "result_id": "finding_918f80084b9f1a35b17535ff1eca7c744a204ed0f8a6150846e09674f3824b7d", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:8502857fb6ed2282", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:5bbb7d9323548b98267932e8a19948e4b98c915aee40a37f47462747ea7c8f5b Applicability: The 33 local configuration boundary cases compared against retained candidate 6 and candidate 7, plus the recorded ten-module integrated suite; no attribution to hidden native assertions. Observation: Candidate 6 fails 31 of 33 configuration boundary cases. Candidate 7 passes all 33 and the full 254-test focused suite. Native history still contains only the two prior submissions and reports cooldown. Interpretation: The tested validation defects are resolved. Integer settings now follow TOML integer types, non-finite numeric limits fail during loading, and special table names are distinguishable in errors. This intentionally rejects integral float count values. API payload validation remains unchanged. Keep candidate 7 for the next eligible evaluation and candidate 6 as the best native-supported recovery source. Continue probing explicit transport registration and the remaining documented contracts; these local results do not prove native improvement or roadmap completion.", + "tags": [ + "writeback-result" + ], + "title": "Does configuration loading reject unusable numeric settings and preserve actionable TOML error paths?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "cd3e40c33a102733", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T19:09:54+00:00", + "result_id": "question_1c2f667dbf653f4505bbdfa5d6c9863a586aa1bc483d6c4bc21f0bb36bb1bf79", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:ef63adcc69a2bb58", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "The 25 local matching cases compared on candidates 7 and 8, including HTTP, WSGI, ASGI and generation, plus the eleven-module suite. No attribution to hidden native assertions.", + "title": "Do explicit serializer registrations take precedence without leaking across vendors or producing wildcard request types?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "fd66bf861399ac38", + "event_kind": "finding", + "evidence_refs": [ + "research/media-matching/REPORT.md", + "research/media-matching/baseline-provenance.json", + "research/media-matching/full-test-results.txt", + "research/media-matching/candidate-8-manifest.json", + "research/media-matching/evaluation-input-manifest.json" + ], + "execution_observation": { + "candidate_sha256": "2ddee2fa6aab4bc24e1830bf03ad3411c8b085b41d4cf24328ae17a19f63ce42", + "change_reason": null, + "elapsed_seconds": 4.55, + "evidence_refs": [ + "research/media-matching/candidate-8-manifest.json", + "research/media-matching/evaluation-input-manifest.json", + "research/media-matching/full-test-results.txt" + ], + "executed_method_id": "pytest.focused-roadmap-suite", + "exit_code": 0, + "input_revision": "sha256:5d39270ffe066f83dce4550f941dc9def28f494e00dbf987e881f99182e91b55", + "limit_seconds": null, + "planned_method_id": "pytest.focused-roadmap-suite", + "result_candidate_sha256": "2ddee2fa6aab4bc24e1830bf03ad3411c8b085b41d4cf24328ae17a19f63ce42", + "result_ref": "research/media-matching/full-test-results.txt", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "All 279 integrated local tests passed. Native evaluation is not yet eligible." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:2ddee2fa6aab4bc24e1830bf03ad3411c8b085b41d4cf24328ae17a19f63ce42", + "comparison_scope": "local:focused-roadmap-suite:sha256:5d39270ffe066f83dce4550f941dc9def28f494e00dbf987e881f99182e91b55", + "evaluation_ref": "pytest:8.4.2:python:3.11:noconftest:roadmap", + "evidence_ref": "research/media-matching/full-test-results.txt", + "higher_is_better": true, + "input_revision": "sha256:5d39270ffe066f83dce4550f941dc9def28f494e00dbf987e881f99182e91b55", + "metric": "focused_tests_passed", + "unit": "tests", + "value": 279 + } + ], + "node_id": "question_1c2f667dbf653f4505bbdfa5d6c9863a586aa1bc483d6c4bc21f0bb36bb1bf79", + "recorded_at": "2026-10-23T19:09:54+00:00", + "result_id": "finding_99ad93ef8a7a0bfdc1656ab56f7a3355b4370902491f8b9bfa0850736df632b8", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:c428d43f1a842bd3", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:5d39270ffe066f83dce4550f941dc9def28f494e00dbf987e881f99182e91b55 Applicability: The 25 local matching cases compared on candidates 7 and 8, including HTTP, WSGI, ASGI and generation, plus the eleven-module suite. No attribution to hidden native assertions. Observation: Candidate 7 fails all 25 new media matching cases. Candidate 8 passes them and the full 279-test focused suite, plus 7 selected existing regressions. Native history contains only the two prior submissions and reports cooldown. Interpretation: The tested selection defects are resolved. Exact matches precede wildcard registrations; more literal characters determine wildcard specificity, with insertion order for ties. Only canonical serializers supply family fallbacks. Schema wildcards expand to concrete registry entries. HTTP verification uses a localhost session isolated from environment proxy settings after the initial default session received 403. Preserve candidate 8 for eligible evaluation and candidate 6 as the best native-supported recovery source. Next audit response factories and remaining metric/check contracts. These local results do not establish a native score improvement or roadmap completion.", + "tags": [ + "writeback-result" + ], + "title": "Do explicit serializer registrations take precedence without leaking across vendors or producing wildcard request types?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "cb2310ef34976fc1", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T19:16:35+00:00", + "result_id": "question_1744b38945b807b582ee3258e54ae4f0b3682b039d0162d762c0a52334096f0f", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:752ba3ed858d93f9", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "The 22 local stream/encoding cases on candidates 8 and 9, plus the twelve-module suite and six existing output tests. Installed clients only; no inference about hidden native assertions.", + "title": "Do response factories normalize streamed content while retaining available upload metadata, timing and encoding behavior?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "795683d21c464182", + "event_kind": "finding", + "evidence_refs": [ + "research/response-streams/REPORT.md", + "research/response-streams/baseline-provenance.json", + "research/response-streams/full-test-results.txt", + "research/response-streams/candidate-9-manifest.json", + "research/response-streams/evaluation-input-manifest.json" + ], + "execution_observation": { + "candidate_sha256": "3e4e14793dd42357f896b3bc0782cedebfdfac9e20449d22128d1c25cd6c2ef4", + "change_reason": null, + "elapsed_seconds": 4.8, + "evidence_refs": [ + "research/response-streams/candidate-9-manifest.json", + "research/response-streams/evaluation-input-manifest.json", + "research/response-streams/full-test-results.txt" + ], + "executed_method_id": "pytest.focused-roadmap-suite", + "exit_code": 0, + "input_revision": "sha256:4a0d5ae4052747cc30a962cfc20abde75325ab42466fef0149e1bd0bae3a0852", + "limit_seconds": null, + "planned_method_id": "pytest.focused-roadmap-suite", + "result_candidate_sha256": "3e4e14793dd42357f896b3bc0782cedebfdfac9e20449d22128d1c25cd6c2ef4", + "result_ref": "research/response-streams/full-test-results.txt", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "All 301 integrated local tests passed. Native evaluation is not yet eligible." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:3e4e14793dd42357f896b3bc0782cedebfdfac9e20449d22128d1c25cd6c2ef4", + "comparison_scope": "local:focused-roadmap-suite:sha256:4a0d5ae4052747cc30a962cfc20abde75325ab42466fef0149e1bd0bae3a0852", + "evaluation_ref": "pytest:8.4.2:python:3.11:noconftest:roadmap", + "evidence_ref": "research/response-streams/full-test-results.txt", + "higher_is_better": true, + "input_revision": "sha256:4a0d5ae4052747cc30a962cfc20abde75325ab42466fef0149e1bd0bae3a0852", + "metric": "focused_tests_passed", + "unit": "tests", + "value": 301 + } + ], + "node_id": "question_1744b38945b807b582ee3258e54ae4f0b3682b039d0162d762c0a52334096f0f", + "recorded_at": "2026-10-23T19:16:35+00:00", + "result_id": "finding_9b3c4a86eb2d151099d89bef5f2d44019e17cb96cf30c9627c56d12d5ee96d2d", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:2182b2d313683f34", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:4a0d5ae4052747cc30a962cfc20abde75325ab42466fef0149e1bd0bae3a0852 Applicability: The 22 local stream/encoding cases on candidates 8 and 9, plus the twelve-module suite and six existing output tests. Installed clients only; no inference about hidden native assertions. Observation: Candidate 8 fails 14 of 22 new stream/encoding cases and passes 8. Candidate 9 passes all 22, all 301 focused tests and 6 existing output tests. Native history contains only the two prior submissions and reports cooldown. Interpretation: The tested stream conversion defects are resolved. HTTPX reads synchronous responses before timing and encoding extraction; streamed uploads retain method, URL and headers with body=None when bytes are unavailable. Closed, interrupted and partially consumed stream errors propagate. Asynchronous bodies require awaiting aread before conversion. Existing Unicode-first JSON detection, charset fallback and cache invalidation remain unchanged, with passing controls. Retain candidate 9 for eligible evaluation and candidate 6 as the best native-supported recovery source. Next audit metric/check selection and runtime configuration. Local results do not establish native improvement or roadmap completion.", + "tags": [ + "writeback-result" + ], + "title": "Do response factories normalize streamed content while retaining available upload metadata, timing and encoding behavior?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "9de669c2bf810ce9", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T19:26:05+00:00", + "result_id": "question_80680bf1f367b484512c3f6b030a87884d480763c6feaa68f9e0dd0ca3038798", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:557c65354fb55593", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "The 16 local extension cases on candidates 9 and 10, thirteen-module focused suite and five existing extension tests. Installed clients only; no inference about hidden native assertions.", + "title": "Are check iterators reusable across operations and are metric names observed once per response?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "6aab781f61139806", + "event_kind": "finding", + "evidence_refs": [ + "research/extension-iteration/REPORT.md", + "research/extension-iteration/baseline-provenance.json", + "research/extension-iteration/full-test-results.txt", + "research/extension-iteration/candidate-10-manifest.json", + "research/extension-iteration/evaluation-input-manifest.json" + ], + "execution_observation": { + "candidate_sha256": "653dc74b886c9f73b11306488d7d7994ad8fdc428b19f958351813162063828b", + "change_reason": null, + "elapsed_seconds": 4.01, + "evidence_refs": [ + "research/extension-iteration/candidate-10-manifest.json", + "research/extension-iteration/evaluation-input-manifest.json", + "research/extension-iteration/full-test-results.txt" + ], + "executed_method_id": "pytest.focused-roadmap-suite", + "exit_code": 0, + "input_revision": "sha256:a371e8d5e61a267f6a89007254c3f19608c86b7104edf2deb29da1f3adc12ec5", + "limit_seconds": null, + "planned_method_id": "pytest.focused-roadmap-suite", + "result_candidate_sha256": "653dc74b886c9f73b11306488d7d7994ad8fdc428b19f958351813162063828b", + "result_ref": "research/extension-iteration/full-test-results.txt", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "All 317 integrated local tests passed. Native evaluation is not yet eligible." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:653dc74b886c9f73b11306488d7d7994ad8fdc428b19f958351813162063828b", + "comparison_scope": "local:focused-roadmap-suite:sha256:a371e8d5e61a267f6a89007254c3f19608c86b7104edf2deb29da1f3adc12ec5", + "evaluation_ref": "pytest:8.4.2:python:3.11:noconftest:roadmap", + "evidence_ref": "research/extension-iteration/full-test-results.txt", + "higher_is_better": true, + "input_revision": "sha256:a371e8d5e61a267f6a89007254c3f19608c86b7104edf2deb29da1f3adc12ec5", + "metric": "focused_tests_passed", + "unit": "tests", + "value": 317 + } + ], + "node_id": "question_80680bf1f367b484512c3f6b030a87884d480763c6feaa68f9e0dd0ca3038798", + "recorded_at": "2026-10-23T19:26:05+00:00", + "result_id": "finding_74f1bf19ed45ea59c582b219561fde11e5f93fa250fb5a90a67feca32b84eb41", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:ff255d98cbe96867", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:a371e8d5e61a267f6a89007254c3f19608c86b7104edf2deb29da1f3adc12ec5 Applicability: The 16 local extension cases on candidates 9 and 10, thirteen-module focused suite and five existing extension tests. Installed clients only; no inference about hidden native assertions. Observation: Candidate 9 fails 14 of 16 new iteration cases and passes 2. Candidate 10 passes all 16, all 317 focused tests and 5 existing extension tests. Native history contains only two prior submissions and reports cooldown. Interpretation: The tested iteration defects are resolved. Runner construction snapshots check iterables for repeated cases, operations and workers while enabled configuration remains effective. Metric collection uses each name once per response; the last callback for a name replaces its predecessor without moving the name. A proposed full Config loader argument was rejected as outside the existing contract after a TypeError probe; no change implements that assumption. Retain candidate 10 for eligible evaluation and candidate 6 as the best native-supported source. Next review configuration discovery and validation against documented contracts and existing tests. Local results do not establish native improvement or roadmap completion.", + "tags": [ + "writeback-result" + ], + "title": "Are check iterators reusable across operations and are metric names observed once per response?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "4299b1f5bcb2bdbe", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T19:32:13+00:00", + "result_id": "question_fe9ebc2f9f8694e71ac1bf331f3a99a1d84f883cd4e7b1291d2b2010de431bf8", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:450502567b3d162b", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Six TOML selections through iterator and generated-test execution, plus four discovery controls on verified candidate 10. Local probe only; no inference about hidden native assertions.", + "title": "Does candidate 10 preserve configured unexpected-method selections through coverage generation and execution?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "074c86bbb01d1ed2", + "event_kind": "finding", + "evidence_refs": [ + "research/replan-coverage/REPORT.md", + "research/replan-coverage/observations.json", + "research/replan-coverage/probe_coverage_configuration.py", + "research/extension-iteration/candidate-10-manifest.json" + ], + "execution_observation": { + "candidate_sha256": "653dc74b886c9f73b11306488d7d7994ad8fdc428b19f958351813162063828b", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "research/extension-iteration/candidate-10-manifest.json", + "research/replan-coverage/observations.json", + "research/replan-coverage/probe_coverage_configuration.py" + ], + "executed_method_id": "python.coverage-configuration-probe", + "exit_code": 0, + "input_revision": "sha256:9df7d456e200ab2be97323e480ec32de9dcf2bfb2ee63c9b68132394d4bc5f98", + "limit_seconds": null, + "planned_method_id": "python.coverage-configuration-probe", + "result_candidate_sha256": "653dc74b886c9f73b11306488d7d7994ad8fdc428b19f958351813162063828b", + "result_ref": "research/replan-coverage/observations.json", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Diagnostic completed: six mismatches and ten matching observations. Next implementation slice selected; source unchanged." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:653dc74b886c9f73b11306488d7d7994ad8fdc428b19f958351813162063828b", + "comparison_scope": "local:coverage-configuration:sha256:9df7d456e200ab2be97323e480ec32de9dcf2bfb2ee63c9b68132394d4bc5f98", + "evaluation_ref": "python:3.11:coverage-configuration-probe:v1", + "evidence_ref": "research/replan-coverage/observations.json", + "higher_is_better": false, + "input_revision": "sha256:9df7d456e200ab2be97323e480ec32de9dcf2bfb2ee63c9b68132394d4bc5f98", + "metric": "contract_mismatches", + "unit": "observations", + "value": 6 + } + ], + "node_id": "question_fe9ebc2f9f8694e71ac1bf331f3a99a1d84f883cd4e7b1291d2b2010de431bf8", + "recorded_at": "2026-10-23T19:32:13+00:00", + "result_id": "finding_cbe3dc433406233f565308a935709ccfb8a8199953ed65493c2111ad3ea30ac2", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:6f86b669c41131f2", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:9df7d456e200ab2be97323e480ec32de9dcf2bfb2ee63c9b68132394d4bc5f98 Applicability: Six TOML selections through iterator and generated-test execution, plus four discovery controls on verified candidate 10. Local probe only; no inference about hidden native assertions. Observation: Candidate 10 produces six mismatches and ten matching observations. Empty method selection falls back to defaults; declared uppercase GET is treated as unspecified; mixed spelling duplicates PATCH. Both execution paths agree. Discovery controls pass. Interpretation: The scoped preservation question has a negative answer. Truthiness fallback and case-sensitive subtraction explain the reproduced execution failures. Preserve this diagnostic and all candidate evidence; prioritize a correction that distinguishes None from empty, normalizes method identity, excludes declared methods and preserves source isolation. Existing uppercase selection usage supports this route. Candidate 10 is unchanged and unsubmitted; candidate 6 remains best native-supported. All six original requirements remain open. Validate the correction under a later implementation admission and match eventual native feedback to the submitted source.", + "tags": [ + "writeback-result" + ], + "title": "Does candidate 10 preserve configured unexpected-method selections through coverage generation and execution?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "ce58416229dfd25b", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T19:37:12+00:00", + "result_id": "question_f9809cac3ad49ec9de31b762c2aa1ef3daeaef852aac333edb946a9027370cbc", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:5d571aa7ec8d5dfa", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "The 24 local coverage cases on candidates 10 and 11, fourteen-module suite and nine existing coverage-example tests. Installed clients only; no inference about hidden native assertions.", + "title": "Do coverage method selections preserve empty lists, exclude declared methods and avoid duplicate generated requests?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "94acde9971f26c83", + "event_kind": "finding", + "evidence_refs": [ + "research/coverage-methods/REPORT.md", + "research/coverage-methods/baseline-provenance.json", + "research/coverage-methods/full-test-results.txt", + "research/coverage-methods/candidate-11-manifest.json", + "research/coverage-methods/evaluation-input-manifest.json" + ], + "execution_observation": { + "candidate_sha256": "b57996f59df3dfaab9ed683e89d8fd97796d205c0cb8fa02cd61a9c6db25234b", + "change_reason": null, + "elapsed_seconds": 5.8, + "evidence_refs": [ + "research/coverage-methods/candidate-11-manifest.json", + "research/coverage-methods/evaluation-input-manifest.json", + "research/coverage-methods/full-test-results.txt" + ], + "executed_method_id": "pytest.focused-roadmap-suite", + "exit_code": 0, + "input_revision": "sha256:d1895aaa886972bde57595ccc5a66d90b2cad9c77886ab34d3f3cc59d98b0db4", + "limit_seconds": null, + "planned_method_id": "pytest.focused-roadmap-suite", + "result_candidate_sha256": "b57996f59df3dfaab9ed683e89d8fd97796d205c0cb8fa02cd61a9c6db25234b", + "result_ref": "research/coverage-methods/full-test-results.txt", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "All 341 integrated local tests passed. Native evaluation is not yet eligible." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:b57996f59df3dfaab9ed683e89d8fd97796d205c0cb8fa02cd61a9c6db25234b", + "comparison_scope": "local:focused-roadmap-suite:sha256:d1895aaa886972bde57595ccc5a66d90b2cad9c77886ab34d3f3cc59d98b0db4", + "evaluation_ref": "pytest:8.4.2:python:3.11:noconftest:roadmap", + "evidence_ref": "research/coverage-methods/full-test-results.txt", + "higher_is_better": true, + "input_revision": "sha256:d1895aaa886972bde57595ccc5a66d90b2cad9c77886ab34d3f3cc59d98b0db4", + "metric": "focused_tests_passed", + "unit": "tests", + "value": 341 + } + ], + "node_id": "question_f9809cac3ad49ec9de31b762c2aa1ef3daeaef852aac333edb946a9027370cbc", + "recorded_at": "2026-10-23T19:37:12+00:00", + "result_id": "finding_8cbf8ad33853da75d17abdf2b57fb2b1d89491ee0b02b0de7ceee59743803222", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:535c5d1c36f89053", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:d1895aaa886972bde57595ccc5a66d90b2cad9c77886ab34d3f3cc59d98b0db4 Applicability: The 24 local coverage cases on candidates 10 and 11, fourteen-module suite and nine existing coverage-example tests. Installed clients only; no inference about hidden native assertions. Observation: Candidate 10 fails 14 of 24 new cases and passes 10. Candidate 11 passes all 24, all 341 focused tests and 9 existing coverage-example tests. Real WSGI/ASGI requests match the selected methods. Native history still reports cooldown. Interpretation: The reproduced selection defects are resolved. None selects defaults; an empty collection disables unspecified-method cases. Normalized method identity excludes declared methods and deduplicates spellings without altering caller configuration. Default HEAD exclusion and explicit HEAD selection remain supported. An initial callback error in the test harness was corrected before the controlled final baseline comparison; both diagnostic logs are retained. Keep candidate 11 for eligible native evaluation and candidate 6 as the best native-supported recovery source. Next audit configured status checks and GraphQL response structure. These local results do not establish native improvement or roadmap completion.", + "tags": [ + "writeback-result" + ], + "title": "Do coverage method selections preserve empty lists, exclude declared methods and avoid duplicate generated requests?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "f374124d0cc20158", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T19:51:08+00:00", + "result_id": "question_29a0b47ce0eeec3d9c59ac406606a653b20c13b93e41129361b068878abe1f92", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:7f2c0af739405be1", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Public native report 889f140e287d matched to candidate-12.zip. Phase totals only; no hidden assertion attribution. Later candidate 13 is a separate unsubmitted correction.", + "title": "Does preserved candidate 12 satisfy all six roadmap phases in native submission 889f140e287d?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "3fa7d397f3a48566", + "event_kind": "finding", + "evidence_refs": [ + "research/check-response-integration/native-result.txt", + "research/check-response-integration/submission-provenance.json", + "research/check-response-integration/candidate-12-manifest.json", + "research/check-response-integration/REPORT.md", + "research/response-buffered/REPORT.md", + "research/response-buffered/baseline-provenance.json", + "research/response-buffered/candidate-6-provenance.json" + ], + "execution_observation": { + "candidate_sha256": "8ee5096d8f0e59b85e9dff2fca792afc298231aa02f66231fee519cc58fd346d", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "research/check-response-integration/candidate-12-manifest.json", + "research/check-response-integration/native-result.txt", + "research/check-response-integration/submission-provenance.json" + ], + "executed_method_id": "sforge-submit.native", + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "limit_seconds": null, + "planned_method_id": "sforge-submit.native", + "result_candidate_sha256": "8ee5096d8f0e59b85e9dff2fca792afc298231aa02f66231fee519cc58fd346d", + "result_ref": "research/check-response-integration/native-result.txt", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Accepted evaluation completed with a lower score; retained its exact source and investigated Response compatibility. All six targets remain open." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:8ee5096d8f0e59b85e9dff2fca792afc298231aa02f66231fee519cc58fd346d", + "comparison_scope": "native:889f140e287d:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/check-response-integration/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "weighted_score", + "unit": "ratio", + "value": 0.775146 + }, + { + "candidate_ref": "sha256:8ee5096d8f0e59b85e9dff2fca792afc298231aa02f66231fee519cc58fd346d", + "comparison_scope": "native:889f140e287d:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/check-response-integration/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "complete_phases", + "unit": "phases", + "value": 1 + }, + { + "candidate_ref": "sha256:8ee5096d8f0e59b85e9dff2fca792afc298231aa02f66231fee519cc58fd346d", + "comparison_scope": "native:889f140e287d:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/check-response-integration/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "configuration_checks_passed", + "unit": "checks", + "value": 19 + }, + { + "candidate_ref": "sha256:8ee5096d8f0e59b85e9dff2fca792afc298231aa02f66231fee519cc58fd346d", + "comparison_scope": "native:889f140e287d:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/check-response-integration/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "api_checks_passed", + "unit": "checks", + "value": 26 + }, + { + "candidate_ref": "sha256:8ee5096d8f0e59b85e9dff2fca792afc298231aa02f66231fee519cc58fd346d", + "comparison_scope": "native:889f140e287d:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/check-response-integration/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "metrics_checks_passed", + "unit": "checks", + "value": 14 + }, + { + "candidate_ref": "sha256:8ee5096d8f0e59b85e9dff2fca792afc298231aa02f66231fee519cc58fd346d", + "comparison_scope": "native:889f140e287d:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/check-response-integration/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "transport_checks_passed", + "unit": "checks", + "value": 14 + }, + { + "candidate_ref": "sha256:8ee5096d8f0e59b85e9dff2fca792afc298231aa02f66231fee519cc58fd346d", + "comparison_scope": "native:889f140e287d:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/check-response-integration/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "response_checks_passed", + "unit": "checks", + "value": 14 + }, + { + "candidate_ref": "sha256:8ee5096d8f0e59b85e9dff2fca792afc298231aa02f66231fee519cc58fd346d", + "comparison_scope": "native:889f140e287d:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/check-response-integration/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "checks_checks_passed", + "unit": "checks", + "value": 16 + } + ], + "node_id": "question_29a0b47ce0eeec3d9c59ac406606a653b20c13b93e41129361b068878abe1f92", + "recorded_at": "2026-10-23T19:51:08+00:00", + "result_id": "finding_e2d44c150745d6dc3b3360c2fa0435948acba701f49dd8a4938c85dba3ce3474", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:ee6cceb2c61591d7", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89 Applicability: Public native report 889f140e287d matched to candidate-12.zip. Phase totals only; no hidden assertion attribution. Later candidate 13 is a separate unsubmitted correction. Observation: Candidate 12 scores 0.775146 with 1/6 complete phases: configuration 19/30, API 26/26, metrics 14/15, transport 14/19, Response 14/18 and checks 16/18. Response is one check below best native candidate 6. Interpretation: The scoped completion question has a negative answer. Preserve candidate 6 as best native-supported source at 0.786257. Reports expose phase totals but no evaluator revision; keep measurement scopes separate and avoid causal attribution. Investigating the Response difference reproduced a buffered-content compatibility regression: three cases fail candidate 12 and pass candidate 6. Later candidate 13 restores cached-content handling, preserves stream behavior and passes 425 focused tests, but is unsubmitted and does not prove score recovery. Candidate 12 also fixes GraphQL configuration/exclusion and detached-response reporting defects; its 81 new cases pass locally, with 71 failures on candidate 11. All snapshots and diagnostic failures remain available. Continue configuration contract review and follow the cooldown before evaluating a later candidate. All six requirements and the Todo remain open.", + "tags": [ + "writeback-result" + ], + "title": "Does preserved candidate 12 satisfy all six roadmap phases in native submission 889f140e287d?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "231a34416c51ae1f", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T19:57:25+00:00", + "result_id": "question_48fcfaadeb864fd28896190931abcd0d54f3aedb4fd9c4b1848305cb411c1aa1", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:93f7fc33177e0e8b", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "The 29 local report configuration cases on candidates 13 and 14, seventeen-module suite and existing CLI reporting property test. Configuration state only; no native assertion attribution.", + "title": "Do selected report formats remain enabled when per-format options are applied while explicit enabled values retain precedence?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "10df5d2b828e6e81", + "event_kind": "finding", + "evidence_refs": [ + "research/report-defaults/REPORT.md", + "research/report-defaults/baseline-provenance.json", + "research/report-defaults/full-test-results.txt", + "research/report-defaults/candidate-14-manifest.json", + "research/report-defaults/evaluation-input-manifest.json" + ], + "execution_observation": { + "candidate_sha256": "aece7422c42250cefdc28cd6070e8bc19684c6f9c97d96f9d134e2d9d90eeadb", + "change_reason": null, + "elapsed_seconds": 6.86, + "evidence_refs": [ + "research/report-defaults/candidate-14-manifest.json", + "research/report-defaults/evaluation-input-manifest.json", + "research/report-defaults/full-test-results.txt" + ], + "executed_method_id": "pytest.focused-roadmap-suite", + "exit_code": 0, + "input_revision": "sha256:c984b3331a035982e49f527cb2b2fb437b1764f31b22149aa6f56b0c5c5cec91", + "limit_seconds": null, + "planned_method_id": "pytest.focused-roadmap-suite", + "result_candidate_sha256": "aece7422c42250cefdc28cd6070e8bc19684c6f9c97d96f9d134e2d9d90eeadb", + "result_ref": "research/report-defaults/full-test-results.txt", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "All 454 integrated local tests passed. Candidate 14 is unsubmitted; native cooldown remains active." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:aece7422c42250cefdc28cd6070e8bc19684c6f9c97d96f9d134e2d9d90eeadb", + "comparison_scope": "local:focused-roadmap-suite:sha256:c984b3331a035982e49f527cb2b2fb437b1764f31b22149aa6f56b0c5c5cec91", + "evaluation_ref": "pytest:8.4.2:python:3.11:noconftest:roadmap", + "evidence_ref": "research/report-defaults/full-test-results.txt", + "higher_is_better": true, + "input_revision": "sha256:c984b3331a035982e49f527cb2b2fb437b1764f31b22149aa6f56b0c5c5cec91", + "metric": "focused_tests_passed", + "unit": "tests", + "value": 454 + } + ], + "node_id": "question_48fcfaadeb864fd28896190931abcd0d54f3aedb4fd9c4b1848305cb411c1aa1", + "recorded_at": "2026-10-23T19:57:25+00:00", + "result_id": "finding_86f44ebabd5ac938d7a6dc429f8b2f5f3836aba07d9b5470a8a2e0c0565ab278", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:4e59f3e6cdbda5e8", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:c984b3331a035982e49f527cb2b2fb437b1764f31b22149aa6f56b0c5c5cec91 Applicability: The 29 local report configuration cases on candidates 13 and 14, seventeen-module suite and existing CLI reporting property test. Configuration state only; no native assertion attribution. Observation: Candidate 13 fails 10 of 29 new cases and passes 19. Candidate 14 passes all 29, all 454 focused tests and the existing CLI reporting property test. Native history still contains only three prior submissions and reports cooldown. Interpretation: The reproduced default-precedence defect is resolved. Selecting a report format supplies enabled=True when an empty or path-only table is constructed; explicit true/false flags and boolean shorthand remain authoritative. Unselected path-only reports remain disabled. Loading through dictionaries, TOML, files and discovery agrees, including environment substitution and isolated input data. This modifies configuration state, not report-writing integration. Preserve candidate 14 for eligible evaluation and candidate 6 as best native-supported. Candidate 12 native feedback remains bound to candidate 12; the retained buffered-response correction has no native recovery proof. Next audit generation defaults and phase/operation overrides. All six requirements remain open; local results do not establish native improvement.", + "tags": [ + "writeback-result" + ], + "title": "Do selected report formats remain enabled when per-format options are applied while explicit enabled values retain precedence?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "320191ab589a8191", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T20:08:08+00:00", + "result_id": "question_08a2cd34c0e8ad00d9cf64a31b8268c6c6055a7bc7c4d2f91510dedec2fabfd0", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:ce10d19c7d038629", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "The 18 local cache and precedence cases on candidates 14 and 15, plus the eighteen-module focused suite. Installed Hypothesis and clients only; no native assertion attribution.", + "title": "Do reused OpenAPI operations apply current character restrictions, codecs and custom header strategies after generation configuration changes?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "a1d5c793baad1870", + "event_kind": "finding", + "evidence_refs": [ + "research/generation-cache/REPORT.md", + "research/generation-cache/baseline-provenance.json", + "research/generation-cache/full-test-results.txt", + "research/generation-cache/candidate-15-manifest.json", + "research/generation-cache/evaluation-input-manifest.json" + ], + "execution_observation": { + "candidate_sha256": "fab3f89ada0be6d70a679664bac721696d52749d0faa589d47cc99fa82752d51", + "change_reason": null, + "elapsed_seconds": 7.83, + "evidence_refs": [ + "research/generation-cache/candidate-15-manifest.json", + "research/generation-cache/evaluation-input-manifest.json", + "research/generation-cache/full-test-results.txt" + ], + "executed_method_id": "pytest.focused-roadmap-suite", + "exit_code": 0, + "input_revision": "sha256:8fa93be7adfec8666e5d0b8e9ae54ce2de080f9318e9aeb7fab2a11cb80e6d55", + "limit_seconds": null, + "planned_method_id": "pytest.focused-roadmap-suite", + "result_candidate_sha256": "fab3f89ada0be6d70a679664bac721696d52749d0faa589d47cc99fa82752d51", + "result_ref": "research/generation-cache/full-test-results.txt", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "All 472 focused tests passed. Candidate 15 is unsubmitted; native cooldown remains active." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:fab3f89ada0be6d70a679664bac721696d52749d0faa589d47cc99fa82752d51", + "comparison_scope": "local:focused-roadmap-suite:sha256:8fa93be7adfec8666e5d0b8e9ae54ce2de080f9318e9aeb7fab2a11cb80e6d55", + "evaluation_ref": "pytest:8.4.2:python:3.11:noconftest:roadmap", + "evidence_ref": "research/generation-cache/full-test-results.txt", + "higher_is_better": true, + "input_revision": "sha256:8fa93be7adfec8666e5d0b8e9ae54ce2de080f9318e9aeb7fab2a11cb80e6d55", + "metric": "focused_tests_passed", + "unit": "tests", + "value": 472 + } + ], + "node_id": "question_08a2cd34c0e8ad00d9cf64a31b8268c6c6055a7bc7c4d2f91510dedec2fabfd0", + "recorded_at": "2026-10-23T20:08:08+00:00", + "result_id": "finding_b7c2895745648accf682b3f1fe69cf4699f92194f05bf8cb48932dfbe68446dd", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:ad634f1f0e3cc4b5", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:8fa93be7adfec8666e5d0b8e9ae54ce2de080f9318e9aeb7fab2a11cb80e6d55 Applicability: The 18 local cache and precedence cases on candidates 14 and 15, plus the eighteen-module focused suite. Installed Hypothesis and clients only; no native assertion attribution. Observation: Candidate 14 fails 15 of 18 final new cases and passes 3. Candidate 15 passes all 18 and all 472 focused tests. Earlier regex probe limitations are retained separately. No new native submission was made. Interpretation: The reproduced stale-cache defect is resolved. Both parameter and body cache keys include allow_x00, codec and the custom header strategy, preserving replacement and in-place edits. Equivalent generation settings reuse entries and cache clearing remains effective. Direct, schema and phase/operation paths agree. Loaded explicit phase defaults pass controls. Direct-constructor propagation and explicit-default provenance are not specified by TODO.md, so their observed differences remain unchanged. Initial regex tests were corrected for installed-library constraints; the final baseline and candidate use identical test bytes. Preserve candidate 15 for eligible native evaluation and candidate 6 as best native-supported. Next audit generation-mode overrides through get_all_tests and direct strategies. All six requirements remain open; local tests do not establish native improvement.", + "tags": [ + "writeback-result" + ], + "title": "Do reused OpenAPI operations apply current character restrictions, codecs and custom header strategies after generation configuration changes?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "b7cb02cabf62ccb4", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T20:18:16+00:00", + "result_id": "question_e118791da2d6af9676a3e21c33bbcbda367692cfdc0c2aa849957fade21c915e", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:3f0aed4bf3da8df6", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "The 46 local mode cases on candidates 15 and 16, revised cache controls and nineteen-module focused suite. OpenAPI payload validity is checked; no native assertion attribution.", + "title": "Do generated tests and direct schema, map and operation strategies honor generation modes while preserving explicit precedence and non-negatable operation behavior?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "baa0197f88a7ae69", + "event_kind": "finding", + "evidence_refs": [ + "research/generation-modes/REPORT.md", + "research/generation-modes/baseline-provenance.json", + "research/generation-modes/full-test-results.txt", + "research/generation-modes/candidate-16-manifest.json", + "research/generation-modes/evaluation-input-manifest.json" + ], + "execution_observation": { + "candidate_sha256": "6d098c6348c31b1bac2c2ec670179c30700a05f199f375a58e6c53a6209f9b57", + "change_reason": null, + "elapsed_seconds": 14.58, + "evidence_refs": [ + "research/generation-modes/candidate-16-manifest.json", + "research/generation-modes/evaluation-input-manifest.json", + "research/generation-modes/full-test-results.txt" + ], + "executed_method_id": "pytest.focused-roadmap-suite", + "exit_code": 0, + "input_revision": "sha256:b1fac54a4da6b3ba4d69738124af662711a64a8738b459d667abb4b3e66b9fd8", + "limit_seconds": null, + "planned_method_id": "pytest.focused-roadmap-suite", + "result_candidate_sha256": "6d098c6348c31b1bac2c2ec670179c30700a05f199f375a58e6c53a6209f9b57", + "result_ref": "research/generation-modes/full-test-results.txt", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "All 518 focused tests passed. Candidate 16 is unsubmitted; native cooldown remains active." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:6d098c6348c31b1bac2c2ec670179c30700a05f199f375a58e6c53a6209f9b57", + "comparison_scope": "local:focused-roadmap-suite:sha256:b1fac54a4da6b3ba4d69738124af662711a64a8738b459d667abb4b3e66b9fd8", + "evaluation_ref": "pytest:8.4.2:python:3.11:noconftest:roadmap", + "evidence_ref": "research/generation-modes/full-test-results.txt", + "higher_is_better": true, + "input_revision": "sha256:b1fac54a4da6b3ba4d69738124af662711a64a8738b459d667abb4b3e66b9fd8", + "metric": "focused_tests_passed", + "unit": "tests", + "value": 518 + } + ], + "node_id": "question_e118791da2d6af9676a3e21c33bbcbda367692cfdc0c2aa849957fade21c915e", + "recorded_at": "2026-10-23T20:18:16+00:00", + "result_id": "finding_54ccbe820152091cf85623399e071f49fab94e35b7b4e8e2cbd489e4ea6e9a2b", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:372a5ba823c66fa8", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:b1fac54a4da6b3ba4d69738124af662711a64a8738b459d667abb4b3e66b9fd8 Applicability: The 46 local mode cases on candidates 15 and 16, revised cache controls and nineteen-module focused suite. OpenAPI payload validity is checked; no native assertion attribution. Observation: Candidate 15 fails 18 of 46 new mode cases and passes 28. Candidate 16 passes all 46 and all 518 focused tests. Revised positive-only cache controls pass all 18 on candidate 15. No native submission was made. Interpretation: The reproduced mode-propagation defect is resolved. Supplied generation modes reach generated tests and direct strategies; explicit single-mode arguments win and omitted configuration retains schema selection. Mixed modes reject an unavailable negative branch rather than skipping a valid positive path; negative-only skips remain. Tests check payload validity, phase overrides, hooks and source isolation. The prior cache tests assumed valid requests despite using a two-mode config; they now request positive mode explicitly and also pass on candidate 15. Original tests and the first integrated failure log remain preserved. Candidate 16 is unsubmitted and candidate 6 remains best native-supported. Next audit security-generation options during operation construction and direct/schema strategy selection. All six requirements remain open; local test totals do not establish native score improvement.", + "tags": [ + "writeback-result" + ], + "title": "Do generated tests and direct schema, map and operation strategies honor generation modes while preserving explicit precedence and non-negatable operation behavior?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "b433f4caf62c55f4", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T20:27:09+00:00", + "result_id": "question_96f5912866c0597da3e560591a0f179550bf3b0844dfd81041b9cfed2cf8bfbe", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:0c6fd785bc9762f2", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "The 74 local security cases on candidates 16 and 17 and twenty-module focused suite. OpenAPI 3 and Swagger 2 generation only; no hidden native assertion attribution.", + "title": "Do security-generation overrides reach operation construction and reused operation strategies while preserving declared parameters and explicit values?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "e9f66439f6e9b771", + "event_kind": "finding", + "evidence_refs": [ + "research/generation-security/REPORT.md", + "research/generation-security/baseline-provenance.json", + "research/generation-security/full-test-results.txt", + "research/generation-security/candidate-17-manifest.json", + "research/generation-security/evaluation-input-manifest.json" + ], + "execution_observation": { + "candidate_sha256": "61a796db601def4a59868f4302212bf02096e2eca376c532f093da5550d7a31d", + "change_reason": null, + "elapsed_seconds": 14.46, + "evidence_refs": [ + "research/generation-security/candidate-17-manifest.json", + "research/generation-security/evaluation-input-manifest.json", + "research/generation-security/full-test-results.txt" + ], + "executed_method_id": "pytest.focused-roadmap-suite", + "exit_code": 0, + "input_revision": "sha256:8afcdac83bb832ffd5436dae0c0d8f02991f86d6c469822360f19b242e96b295", + "limit_seconds": null, + "planned_method_id": "pytest.focused-roadmap-suite", + "result_candidate_sha256": "61a796db601def4a59868f4302212bf02096e2eca376c532f093da5550d7a31d", + "result_ref": "research/generation-security/full-test-results.txt", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "All 592 focused tests passed. Candidate 17 is unsubmitted; native cooldown remains active." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:61a796db601def4a59868f4302212bf02096e2eca376c532f093da5550d7a31d", + "comparison_scope": "local:focused-roadmap-suite:sha256:8afcdac83bb832ffd5436dae0c0d8f02991f86d6c469822360f19b242e96b295", + "evaluation_ref": "pytest:8.4.2:python:3.11:noconftest:roadmap", + "evidence_ref": "research/generation-security/full-test-results.txt", + "higher_is_better": true, + "input_revision": "sha256:8afcdac83bb832ffd5436dae0c0d8f02991f86d6c469822360f19b242e96b295", + "metric": "focused_tests_passed", + "unit": "tests", + "value": 592 + } + ], + "node_id": "question_96f5912866c0597da3e560591a0f179550bf3b0844dfd81041b9cfed2cf8bfbe", + "recorded_at": "2026-10-23T20:27:09+00:00", + "result_id": "finding_64e63cb60eddea8ced76eaf4b51e3d70c9132abb33ec8095d8fae5f533fc8f27", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:0f8ce616d4e1f278", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:8afcdac83bb832ffd5436dae0c0d8f02991f86d6c469822360f19b242e96b295 Applicability: The 74 local security cases on candidates 16 and 17 and twenty-module focused suite. OpenAPI 3 and Swagger 2 generation only; no hidden native assertion attribution. Observation: Candidate 16 fails 47 of 74 new cases and passes 27. Candidate 17 passes all 74 and all 592 focused tests. Explicit parameters, values, local security removal and negative-only skips remain effective. No native submission was made. Interpretation: The reproduced security-setting defects are resolved. Schema strategy construction forwards the override. Reused operations generate from the effective parameter set, cache by security setting and serialize that set; negative eligibility uses it too. Synthesized security parameters are distinguished from declared parameters without mutating the operation. HTTP authentication preserves declared Authorization inputs. Both override directions, schema mutations, repeated reuse, generated tests, hooks and Swagger are covered. The initial 60-case comparison and final expanded baseline remain preserved. Candidate 17 is unsubmitted and candidate 6 remains best native-supported. Next audit GraphQL generation settings at public entry points and check eligibility before evaluation. All six requirements remain open; local totals do not establish native improvement.", + "tags": [ + "writeback-result" + ], + "title": "Do security-generation overrides reach operation construction and reused operation strategies while preserving declared parameters and explicit values?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "83d9b77eff6cc560", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T20:33:59+00:00", + "result_id": "question_469d2c2556ea929d1647b6972010536d55d7d2e984798fe3748fd82da8adccd6", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:6099333ac6903f22", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Candidate 17: eight 20-example configuration paths, three permissive controls and lookup/enumeration/execution comparisons across two access orders. Local GraphQL only; no native inference.", + "title": "Do GraphQL generation settings and cached field lookups preserve requested constraints and query or mutation root identity?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "0136b62dacd6729f", + "event_kind": "finding", + "evidence_refs": [ + "research/replan-graphql/REPORT.md", + "research/replan-graphql/observations.json", + "research/replan-graphql/probe_graphql.py", + "research/generation-security/candidate-17-manifest.json", + "research/replan-graphql/execution.json" + ], + "execution_observation": { + "candidate_sha256": "61a796db601def4a59868f4302212bf02096e2eca376c532f093da5550d7a31d", + "change_reason": null, + "elapsed_seconds": 2.6450722929876065, + "evidence_refs": [ + "research/generation-security/candidate-17-manifest.json", + "research/replan-graphql/execution.json", + "research/replan-graphql/observations.json", + "research/replan-graphql/probe_graphql.py" + ], + "executed_method_id": "python.graphql-generation-root-probe", + "exit_code": 0, + "input_revision": "sha256:8e31366cc1281d899e0978d699792cfd37af5b690bb37f467165619439db9852", + "limit_seconds": null, + "planned_method_id": "python.graphql-generation-root-probe", + "result_candidate_sha256": "61a796db601def4a59868f4302212bf02096e2eca376c532f093da5550d7a31d", + "result_ref": "research/replan-graphql/observations.json", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Bounded review completed with ten mismatches and 53 controls. Root cache correction selected; implementation unchanged." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:61a796db601def4a59868f4302212bf02096e2eca376c532f093da5550d7a31d", + "comparison_scope": "local:graphql-generation-root-identity:sha256:8e31366cc1281d899e0978d699792cfd37af5b690bb37f467165619439db9852", + "evaluation_ref": "python:3.11:graphql-generation-root-probe:v1", + "evidence_ref": "research/replan-graphql/observations.json", + "higher_is_better": false, + "input_revision": "sha256:8e31366cc1281d899e0978d699792cfd37af5b690bb37f467165619439db9852", + "metric": "contract_mismatches", + "unit": "observations", + "value": 10 + } + ], + "node_id": "question_469d2c2556ea929d1647b6972010536d55d7d2e984798fe3748fd82da8adccd6", + "recorded_at": "2026-10-23T20:33:59+00:00", + "result_id": "finding_3e3e5c42b01eee73953b80f237b1e7972f2b43370a9552bec080f0cf300834d5", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:677a46742b17b810", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:8e31366cc1281d899e0978d699792cfd37af5b690bb37f467165619439db9852 Applicability: Candidate 17: eight 20-example configuration paths, three permissive controls and lookup/enumeration/execution comparisons across two access orders. Local GraphQL only; no native inference. Observation: The final probe records ten mismatches and 53 matching observations, with no unknowns. Generation restrictions pass. Shared field names return the first cached root, and local execution confirms wrong query/mutation routing. Source is unchanged. Interpretation: The scoped preservation question has a negative answer. Null, NUL and codec restrictions pass across loader/call configuration and schema/map/operation/generated-test paths with source isolation. The new defect is field-only operation caching: Query.lookup and Mutation.lookup share one object. Both lookup orders fail on the second root through direct and map strategies; execution uses the wrong root resolver. Distinct names, first lookups, enumeration and generated tests pass. Prioritize root-aware cache identity and retain repeat-lookup reuse, custom root names, independent schemas and configuration controls in regression validation. The original objective and all six requirements remain accurate and open. Candidate 17 is unchanged and unsubmitted; candidate 6 remains native best. Preserve the initial probe separately and follow native cooldown before evaluating a later source.", + "tags": [ + "writeback-result" + ], + "title": "Do GraphQL generation settings and cached field lookups preserve requested constraints and query or mutation root identity?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "aa8c5af707dd75f3", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T20:40:50+00:00", + "result_id": "question_e345e5b8630e7c7e61af7ce0f50b73f0afaa6184bf59c67ef55a44e22f84e4de", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:cabc9c5a35b910ab", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "The 38 final local GraphQL cases on candidates 17 and 18, 21-module focused suite and three existing scalar tests. No attribution to hidden native assertions.", + "title": "Do root-qualified GraphQL cache keys preserve requested operation identity and execution while retaining lookup errors and generation controls?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "323f7bb3d6fe8ed7", + "event_kind": "finding", + "evidence_refs": [ + "research/graphql-cache/REPORT.md", + "research/graphql-cache/baseline-provenance.json", + "research/graphql-cache/full-test-results.txt", + "research/graphql-cache/candidate-18-manifest.json", + "research/graphql-cache/evaluation-input-manifest.json" + ], + "execution_observation": { + "candidate_sha256": "d80c44537952fdee4137362417c4416a464854771029ae2d9b3354dbb632b98f", + "change_reason": null, + "elapsed_seconds": 18.02, + "evidence_refs": [ + "research/graphql-cache/candidate-18-manifest.json", + "research/graphql-cache/evaluation-input-manifest.json", + "research/graphql-cache/full-test-results.txt" + ], + "executed_method_id": "pytest.focused-roadmap-suite", + "exit_code": 0, + "input_revision": "sha256:76f588aa60b83572b20110d6e4133a6f51d0c56c72bf288f8fef8196d9db1609", + "limit_seconds": null, + "planned_method_id": "pytest.focused-roadmap-suite", + "result_candidate_sha256": "d80c44537952fdee4137362417c4416a464854771029ae2d9b3354dbb632b98f", + "result_ref": "research/graphql-cache/full-test-results.txt", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "All 630 focused tests passed. Candidate 18 is unsubmitted; native cooldown remains active." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:d80c44537952fdee4137362417c4416a464854771029ae2d9b3354dbb632b98f", + "comparison_scope": "local:focused-roadmap-suite:sha256:76f588aa60b83572b20110d6e4133a6f51d0c56c72bf288f8fef8196d9db1609", + "evaluation_ref": "pytest:8.4.2:python:3.11:noconftest:roadmap", + "evidence_ref": "research/graphql-cache/full-test-results.txt", + "higher_is_better": true, + "input_revision": "sha256:76f588aa60b83572b20110d6e4133a6f51d0c56c72bf288f8fef8196d9db1609", + "metric": "focused_tests_passed", + "unit": "tests", + "value": 630 + } + ], + "node_id": "question_e345e5b8630e7c7e61af7ce0f50b73f0afaa6184bf59c67ef55a44e22f84e4de", + "recorded_at": "2026-10-23T20:40:50+00:00", + "result_id": "finding_1012e1bbec0a648ae9420ae56e7f9383dbb64412a8f148b14c10988493559872", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:e659f398550bca32", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:76f588aa60b83572b20110d6e4133a6f51d0c56c72bf288f8fef8196d9db1609 Applicability: The 38 final local GraphQL cases on candidates 17 and 18, 21-module focused suite and three existing scalar tests. No attribution to hidden native assertions. Observation: Candidate 17 fails 14 of 38 final new cases and passes 24. Candidate 18 passes all 38, all 630 focused tests and three existing GraphQL tests. Server history still reports cooldown; no submission was attempted. Interpretation: The reproduced root collision is resolved. Cache keys include the root type name and field name, preserving repeated lookup identity within a root. Direct/map strategies execute the requested query or mutation, including custom root names and both access orders. Missing fields on one root are no longer supplied from another root cache entry. Independent schemas, enumeration, generated tests and generation restrictions pass controls. The initial NUL reachability search exhausted its budget on both candidates; its evidence remains separate. Final comparison uses identical inputs with a fixed-seed 1000-example reachability budget. Candidate 18 remains unsubmitted and candidate 6 remains best native-supported. Recheck eligibility before evaluation; meanwhile review remaining Registry and metric contracts. Local success does not establish native score recovery or roadmap completion.", + "tags": [ + "writeback-result" + ], + "title": "Do root-qualified GraphQL cache keys preserve requested operation identity and execution while retaining lookup errors and generation controls?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "eaf937418c4ef89e", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T20:51:14+00:00", + "result_id": "question_33693f251630347ff993be73bf6d07221caeb3e20fc1eb6152277ca4102b6317", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:eafe90f6be228b96", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Public native report 0aac8a5d5809 matched to candidate-18.zip. Phase totals only; no hidden assertion attribution. The separate registry/metric audit leaves this source unchanged.", + "title": "Does preserved candidate 18 satisfy all six roadmap phases in native submission 0aac8a5d5809?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "6dc9ea09a602c29b", + "event_kind": "finding", + "evidence_refs": [ + "research/registry-metrics-audit/REPORT.md", + "research/registry-metrics-audit/native-result.txt", + "research/registry-metrics-audit/submission-source.json", + "research/registry-metrics-audit/focus-correction.md", + "research/registry-metrics-audit/observations.json", + "research/registry-metrics-audit/acceptance-review.md", + "research/graphql-cache/candidate-18-manifest.json" + ], + "execution_observation": { + "candidate_sha256": "d80c44537952fdee4137362417c4416a464854771029ae2d9b3354dbb632b98f", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "research/graphql-cache/candidate-18-manifest.json", + "research/registry-metrics-audit/native-result.txt", + "research/registry-metrics-audit/submission-source.json" + ], + "executed_method_id": "sforge-submit.native", + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "limit_seconds": null, + "planned_method_id": "sforge-submit.native", + "result_candidate_sha256": "d80c44537952fdee4137362417c4416a464854771029ae2d9b3354dbb632b98f", + "result_ref": "research/registry-metrics-audit/native-result.txt", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Accepted native evaluation completed with a best-score tie and five incomplete phases. Original scope stays open." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:d80c44537952fdee4137362417c4416a464854771029ae2d9b3354dbb632b98f", + "comparison_scope": "native:0aac8a5d5809:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/registry-metrics-audit/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "weighted_score", + "unit": "ratio", + "value": 0.786257 + }, + { + "candidate_ref": "sha256:d80c44537952fdee4137362417c4416a464854771029ae2d9b3354dbb632b98f", + "comparison_scope": "native:0aac8a5d5809:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/registry-metrics-audit/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "complete_phases", + "unit": "phases", + "value": 1 + }, + { + "candidate_ref": "sha256:d80c44537952fdee4137362417c4416a464854771029ae2d9b3354dbb632b98f", + "comparison_scope": "native:0aac8a5d5809:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/registry-metrics-audit/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "configuration_checks_passed", + "unit": "checks", + "value": 19 + }, + { + "candidate_ref": "sha256:d80c44537952fdee4137362417c4416a464854771029ae2d9b3354dbb632b98f", + "comparison_scope": "native:0aac8a5d5809:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/registry-metrics-audit/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "api_checks_passed", + "unit": "checks", + "value": 26 + }, + { + "candidate_ref": "sha256:d80c44537952fdee4137362417c4416a464854771029ae2d9b3354dbb632b98f", + "comparison_scope": "native:0aac8a5d5809:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/registry-metrics-audit/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "metrics_checks_passed", + "unit": "checks", + "value": 14 + }, + { + "candidate_ref": "sha256:d80c44537952fdee4137362417c4416a464854771029ae2d9b3354dbb632b98f", + "comparison_scope": "native:0aac8a5d5809:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/registry-metrics-audit/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "transport_checks_passed", + "unit": "checks", + "value": 14 + }, + { + "candidate_ref": "sha256:d80c44537952fdee4137362417c4416a464854771029ae2d9b3354dbb632b98f", + "comparison_scope": "native:0aac8a5d5809:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/registry-metrics-audit/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "response_checks_passed", + "unit": "checks", + "value": 15 + }, + { + "candidate_ref": "sha256:d80c44537952fdee4137362417c4416a464854771029ae2d9b3354dbb632b98f", + "comparison_scope": "native:0aac8a5d5809:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/registry-metrics-audit/native-result.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "checks_checks_passed", + "unit": "checks", + "value": 16 + } + ], + "node_id": "question_33693f251630347ff993be73bf6d07221caeb3e20fc1eb6152277ca4102b6317", + "recorded_at": "2026-10-23T20:51:14+00:00", + "result_id": "finding_ca36eade494ed8ed91e951338f43b2f1704f467990da36ec884bc09b1e75b6d9", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:a29e424fbaa5d8fb", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89 Applicability: Public native report 0aac8a5d5809 matched to candidate-18.zip. Phase totals only; no hidden assertion attribution. The separate registry/metric audit leaves this source unchanged. Observation: Candidate 18 scores 0.786257 with 1/6 phases complete: configuration 19/30, API 26/26, metrics 14/15, transport 14/19, Response 15/18, checks 16/18. It ties candidate 6; Response recovers one reported check from candidate 12. Interpretation: The scoped completion question has a negative answer. Retain candidates 6 and 18 as tied native best; candidate 18 is current. All 204 captured files matched before and after submission. One accepted evaluation was queried by its exact ID; the first pending result was retained. Public reports do not disclose evaluator revision or individual assertions, so scopes stay separate and recovery is not attributed to a specific patch. This wake also records 33 matching registry/metric observations and no mismatches; the prior 630-case suite was not rerun on unchanged source. Local coverage expansion did not close the remaining phase gaps. The acceptance mapping and focus correction prioritize numbered configuration contracts and boundary behavior rather than more incidental runtime fixes. All six requirements and the integration Todo remain open; honor the renewed cooldown.", + "tags": [ + "writeback-result" + ], + "title": "Does preserved candidate 18 satisfy all six roadmap phases in native submission 0aac8a5d5809?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "58131b5cc6ad1073", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T21:02:11+00:00", + "result_id": "question_c0c31fb04d756b040a482d95049fb8b0f786dc65ac70fe379e3977729466d3df", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:1607b16adc9fe48c", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "The 23 local selection cases on candidates 18 and 19 and the 22-module focused suite. Installed clients only; no hidden native assertion attribution.", + "title": "Does transport selection preserve asynchronous callable behavior through partial applications while retaining synchronous selection and actual execution?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "375fc06b51dc50a1", + "event_kind": "finding", + "evidence_refs": [ + "research/transport-selection/REPORT.md", + "research/transport-selection/baseline-provenance.json", + "research/transport-selection/full-test-results.txt", + "research/transport-selection/candidate-19-manifest.json", + "research/transport-selection/evaluation-input-manifest.json", + "research/config-contract-audit/REPORT.md" + ], + "execution_observation": { + "candidate_sha256": "46c22159c581602899e992bee6b69744fd475795378ec5e35dacf1afbc01f139", + "change_reason": null, + "elapsed_seconds": 20.83, + "evidence_refs": [ + "research/transport-selection/candidate-19-manifest.json", + "research/transport-selection/evaluation-input-manifest.json", + "research/transport-selection/full-test-results.txt" + ], + "executed_method_id": "pytest.focused-roadmap-suite", + "exit_code": 0, + "input_revision": "sha256:aa942e1284123170e8b0be09068e27f6b94c2a208ac2f11816b66f2c047ecd06", + "limit_seconds": null, + "planned_method_id": "pytest.focused-roadmap-suite", + "result_candidate_sha256": "46c22159c581602899e992bee6b69744fd475795378ec5e35dacf1afbc01f139", + "result_ref": "research/transport-selection/full-test-results.txt", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "All 653 focused tests passed. Candidate 19 is unsubmitted; native cooldown remains active." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:46c22159c581602899e992bee6b69744fd475795378ec5e35dacf1afbc01f139", + "comparison_scope": "local:focused-roadmap-suite:sha256:aa942e1284123170e8b0be09068e27f6b94c2a208ac2f11816b66f2c047ecd06", + "evaluation_ref": "pytest:8.4.2:python:3.11:noconftest:roadmap", + "evidence_ref": "research/transport-selection/full-test-results.txt", + "higher_is_better": true, + "input_revision": "sha256:aa942e1284123170e8b0be09068e27f6b94c2a208ac2f11816b66f2c047ecd06", + "metric": "focused_tests_passed", + "unit": "tests", + "value": 653 + } + ], + "node_id": "question_c0c31fb04d756b040a482d95049fb8b0f786dc65ac70fe379e3977729466d3df", + "recorded_at": "2026-10-23T21:02:11+00:00", + "result_id": "finding_c5c8b734810a18de64194680b5cb568cf639b61e8ae4df4a0aa7a7f74f0019a3", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:5c763d6682f52bb2", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:aa942e1284123170e8b0be09068e27f6b94c2a208ac2f11816b66f2c047ecd06 Applicability: The 23 local selection cases on candidates 18 and 19 and the 22-module focused suite. Installed clients only; no hidden native assertion attribution. Observation: Candidate 18 fails 6 of 23 selection cases and passes 17. Candidate 19 passes all 23 and all 653 focused tests. Wrapped async applications execute correctly, synchronous controls remain WSGI and bound arguments are retained. Interpretation: The reproduced Target 4 defect is resolved. inspect.iscoroutinefunction handles partial functions but not partials of async callable objects; the helper now inspects their wrapped callable without invoking or replacing the application. Actual Starlette requests no longer receive the WSGI argument list. Source and input manifests retain the controlled comparison. Before this correction, a separate 51-observation configuration audit on unchanged candidate 18 found no mismatch, so no configuration defaults or aliases changed. Candidate 19 is unsubmitted; candidates 6 and 18 remain tied native best. Local test counts do not establish native improvement or full completion. Continue direct Response contract review and honor eligibility before another submission; all six requirements and the existing Todo remain open.", + "tags": [ + "writeback-result" + ], + "title": "Does transport selection preserve asynchronous callable behavior through partial applications while retaining synchronous selection and actual execution?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "192c5d9f3b9869ba", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T21:12:44+00:00", + "result_id": "question_080b0c13f68b7fb2bdca7546ffe5fe8831119cb535f589635035c38d1cd37270", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:c1eb4b73ea460dd0", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "The 26 local envelope cases on candidates 19 and 20 and the 23-module focused suite. Direct checks and WSGI/ASGI calls; no hidden native assertion attribution.", + "title": "Does GraphQL response validation reject unrecognized envelope fields while preserving application data, extensions and valid error classifications?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "2e8410625d7678ce", + "event_kind": "finding", + "evidence_refs": [ + "research/graphql-response-shape/REPORT.md", + "research/graphql-response-shape/baseline-provenance.json", + "research/graphql-response-shape/full-test-results.txt", + "research/graphql-response-shape/candidate-20-manifest.json", + "research/graphql-response-shape/evaluation-input-manifest.json", + "research/response-contract-audit/REPORT.md", + "research/graphql-response-shape/existing-regression.json" + ], + "execution_observation": { + "candidate_sha256": "9e3d607d68aee887be7cc06b4c5463b248346645f630f484bf0a290a152ddfd1", + "change_reason": null, + "elapsed_seconds": 17.92, + "evidence_refs": [ + "research/graphql-response-shape/candidate-20-manifest.json", + "research/graphql-response-shape/evaluation-input-manifest.json", + "research/graphql-response-shape/full-test-results.txt" + ], + "executed_method_id": "pytest.focused-roadmap-suite", + "exit_code": 0, + "input_revision": "sha256:629787890540e817736c3937ea73fe3f7ab006f2e21e1478c4c67d65492f2e56", + "limit_seconds": null, + "planned_method_id": "pytest.focused-roadmap-suite", + "result_candidate_sha256": "9e3d607d68aee887be7cc06b4c5463b248346645f630f484bf0a290a152ddfd1", + "result_ref": "research/graphql-response-shape/full-test-results.txt", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "All 679 focused tests passed. Candidate 20 is unsubmitted; native cooldown remains active." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:9e3d607d68aee887be7cc06b4c5463b248346645f630f484bf0a290a152ddfd1", + "comparison_scope": "local:focused-roadmap-suite:sha256:629787890540e817736c3937ea73fe3f7ab006f2e21e1478c4c67d65492f2e56", + "evaluation_ref": "pytest:8.4.2:python:3.11:noconftest:roadmap", + "evidence_ref": "research/graphql-response-shape/full-test-results.txt", + "higher_is_better": true, + "input_revision": "sha256:629787890540e817736c3937ea73fe3f7ab006f2e21e1478c4c67d65492f2e56", + "metric": "focused_tests_passed", + "unit": "tests", + "value": 679 + } + ], + "node_id": "question_080b0c13f68b7fb2bdca7546ffe5fe8831119cb535f589635035c38d1cd37270", + "recorded_at": "2026-10-23T21:12:44+00:00", + "result_id": "finding_8015ab187e747e21b2c15a9e0519c7ab613c3bb63b597ecccfeac6bef5fc75e7", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:7610ab3f45f65ba9", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:629787890540e817736c3937ea73fe3f7ab006f2e21e1478c4c67d65492f2e56 Applicability: The 26 local envelope cases on candidates 19 and 20 and the 23-module focused suite. Direct checks and WSGI/ASGI calls; no hidden native assertion attribution. Observation: Candidate 19 fails 14 of 26 envelope cases and passes 12. Candidate 20 passes all 26 and all 679 focused tests. Extra envelope fields produce structural failures; data, extensions and valid error classifications remain supported. Interpretation: The reproduced Target 6 structural gap is resolved. Only data, errors and extensions are accepted as top-level GraphQL response fields. Arbitrary nested data and extension keys remain valid, and per-error metadata is unchanged. The exact source/test comparison and candidate archive are preserved. One unchanged legacy multiple-error test body passes in isolation; its module imports a removed v3 path, so full legacy collection is not claimed. A separate prior 61-observation Response audit on unchanged candidate 19 found no mismatch and justified no Response change. Candidate 20 is unsubmitted; candidates 6 and 18 remain tied native best. Local coverage does not prove native improvement or completion. Review direct serializer registration consistency next, honoring cooldown and retaining every original requirement and the open Todo.", + "tags": [ + "writeback-result" + ], + "title": "Does GraphQL response validation reject unrecognized envelope fields while preserving application data, extensions and valid error classifications?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "c6b9dcdc0b4049f4", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T21:21:02+00:00", + "result_id": "question_b004aba42c3592a1835216e49b3740bd9fffc6e144bf2fea321e3a578ddee1b3", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:ac82ed156d5bf1ca", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "The 49 public lifecycle observations on unchanged candidate 20, including actual local HTTP/WSGI/ASGI calls. No hidden native assertion attribution or concurrent registry mutation guarantee.", + "title": "Do serializer registration, replacement and removal remain consistent across reused cases, generation strategies and HTTP, WSGI and ASGI execution?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "6e75edc20ddd1c1c", + "event_kind": "finding", + "evidence_refs": [ + "research/serializer-lifecycle/REPORT.md", + "research/serializer-lifecycle/observations.json", + "research/serializer-lifecycle/probe.py", + "research/serializer-lifecycle/execution.json", + "research/graphql-response-shape/candidate-20-manifest.json" + ], + "execution_observation": { + "candidate_sha256": "9e3d607d68aee887be7cc06b4c5463b248346645f630f484bf0a290a152ddfd1", + "change_reason": null, + "elapsed_seconds": 1.1549159589922056, + "evidence_refs": [ + "research/graphql-response-shape/candidate-20-manifest.json", + "research/serializer-lifecycle/execution.json", + "research/serializer-lifecycle/observations.json", + "research/serializer-lifecycle/probe.py" + ], + "executed_method_id": "python.serializer-lifecycle-probe", + "exit_code": 0, + "input_revision": "sha256:ad1fbeef4d3658cbec66d5b854debabe107617bf48f5b639832ba821de68fa05", + "limit_seconds": null, + "planned_method_id": "python.serializer-lifecycle-probe", + "result_candidate_sha256": "9e3d607d68aee887be7cc06b4c5463b248346645f630f484bf0a290a152ddfd1", + "result_ref": "research/serializer-lifecycle/observations.json", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Bounded audit completed with 49 matches and no mismatches. Source remains unchanged; all six roadmap requirements remain open." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:9e3d607d68aee887be7cc06b4c5463b248346645f630f484bf0a290a152ddfd1", + "comparison_scope": "local:serializer-lifecycle:sha256:ad1fbeef4d3658cbec66d5b854debabe107617bf48f5b639832ba821de68fa05", + "evaluation_ref": "python:3.11:serializer-lifecycle-probe:v1", + "evidence_ref": "research/serializer-lifecycle/observations.json", + "higher_is_better": false, + "input_revision": "sha256:ad1fbeef4d3658cbec66d5b854debabe107617bf48f5b639832ba821de68fa05", + "metric": "contract_mismatches", + "unit": "observations", + "value": 0 + } + ], + "node_id": "question_b004aba42c3592a1835216e49b3740bd9fffc6e144bf2fea321e3a578ddee1b3", + "recorded_at": "2026-10-23T21:21:02+00:00", + "result_id": "finding_42a7f6ff0d870f2bdf35a7abddbaa3edebbfde504393954b0e5d8b4757c78450", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:95e29b1cd34170d2", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:ad1fbeef4d3658cbec66d5b854debabe107617bf48f5b639832ba821de68fa05 Applicability: The 49 public lifecycle observations on unchanged candidate 20, including actual local HTTP/WSGI/ASGI calls. No hidden native assertion attribution or concurrent registry mutation guarantee. Observation: Candidate 20 matches all 49 observations with zero mismatches or unknowns. Reused cases and generation see registry changes; all three transports send replacement bytes, omit bodies and recover after re-registration. All 206 candidate files remain unchanged. Interpretation: The scoped consistency question has a positive answer. Function/context identity, alias independence, transport isolation, replacement, removal, callback errors and wire body transitions pass. The suspected lifetime defect was not reproduced, so no production change is justified. Existing media matching results remain applicable and were not rerun. The probe restores registry state and verifies the complete candidate archive before and after execution. Candidate 20 remains unsubmitted; candidates 6 and 18 remain native best with only API complete. This negative diagnostic does not explain the remaining transport checks or close the roadmap. Next review direct GenerationConfig option conversion through phase/project loading, using existing evidence and requiring a public reproduction. Follow native cooldown and preserve the original requirements and open Todo.", + "tags": [ + "writeback-result" + ], + "title": "Do serializer registration, replacement and removal remain consistent across reused cases, generation strategies and HTTP, WSGI and ASGI execution?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "4572257dacce6bef", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T21:36:32+00:00", + "result_id": "question_a4d444b314f9ea538eae0fcd49d0fbd819ed88b912a79d56bc93394996b1a551", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:d1b4dd70795588d0", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Ten changed guides, 119 Python blocks and 39 targeted observations on candidate 21. Selected extension examples execute locally; no rendered docs or external application examples are claimed.", + "title": "Do migrated active guides and missing-URL guidance use the implemented v4 API and executable serializer, metric and check examples?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "4ed1866911f56239", + "event_kind": "finding", + "evidence_refs": [ + "research/api-guides/REPORT.md", + "research/api-guides/validation-results.json", + "research/api-guides/execution.json", + "research/api-guides/candidate-21-manifest.json", + "research/api-guides/evaluation-input-manifest.json", + "research/generation-config-audit/REPORT.md" + ], + "execution_observation": { + "candidate_sha256": "e36d69cc5f7d5fe8a063fa7d6289256b2c36bc1413e20840f1d64614068844b4", + "change_reason": null, + "elapsed_seconds": 0.9850239590014098, + "evidence_refs": [ + "research/api-guides/candidate-21-manifest.json", + "research/api-guides/evaluation-input-manifest.json", + "research/api-guides/execution.json", + "research/api-guides/validation-results.json" + ], + "executed_method_id": "python.api-guide-validation", + "exit_code": 0, + "input_revision": "sha256:b66fb0275dbda4180829fd5c29475e12de31a7386e89da71dd61982924d70262", + "limit_seconds": null, + "planned_method_id": "python.api-guide-validation", + "result_candidate_sha256": "e36d69cc5f7d5fe8a063fa7d6289256b2c36bc1413e20840f1d64614068844b4", + "result_ref": "research/api-guides/validation-results.json", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "All 39 observations pass and 119 Python blocks parse. Sphinx is unavailable. Candidate 21 is unsubmitted." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:e36d69cc5f7d5fe8a063fa7d6289256b2c36bc1413e20840f1d64614068844b4", + "comparison_scope": "local:api-guide-validation:sha256:b66fb0275dbda4180829fd5c29475e12de31a7386e89da71dd61982924d70262", + "evaluation_ref": "python:3.11:api-guide-validation:v1", + "evidence_ref": "research/api-guides/validation-results.json", + "higher_is_better": true, + "input_revision": "sha256:b66fb0275dbda4180829fd5c29475e12de31a7386e89da71dd61982924d70262", + "metric": "guide_validation_observations_passed", + "unit": "observations", + "value": 39 + } + ], + "node_id": "question_a4d444b314f9ea538eae0fcd49d0fbd819ed88b912a79d56bc93394996b1a551", + "recorded_at": "2026-10-23T21:36:32+00:00", + "result_id": "finding_43fb561f5d156ca686770130757de77be862fa1f086f1088c90266c2deb39dd0", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:e8ddbf77d6999cd4", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:b66fb0275dbda4180829fd5c29475e12de31a7386e89da71dd61982924d70262 Applicability: Ten changed guides, 119 Python blocks and 39 targeted observations on candidate 21. Selected extension examples execute locally; no rendered docs or external application examples are claimed. Observation: Candidate 21 passes all 39 observations and all 119 guide Python blocks parse. The active-guide scan drops from 61 obsolete references to zero. Documented CSV, metric and check examples execute locally. URL guidance uses current namespaces. Interpretation: The scoped guide migration is validated. Loaders, configuration imports, generation modes, unified Response, metric registration and three-argument checks use current APIs. CSV serialization uses one bytes-returning function across transports. Original documents and the source-message baseline are preserved. One incomplete Python block is corrected. Remote URL examples and app-specific fixtures were not executed; Sphinx/docutils are unavailable, so rendering is unverified. The earlier configuration audit matches 87 observations on unchanged candidate 20 and justified no config change. Candidate 21 remains unsubmitted; native best is the 6/18 tie with only API complete. Documentation success is not native improvement or roadmap completion. Keep all six requirements and the same open Todo; validate offline packaging with installed tooling and evaluate when eligible.", + "tags": [ + "writeback-result" + ], + "title": "Do migrated active guides and missing-URL guidance use the implemented v4 API and executable serializer, metric and check examples?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "b580393d781216bc", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T21:43:20+00:00", + "result_id": "question_dbc149db90f786f07c47493d05de4541b670ad9e055813f7cf1ef7de8789148e", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:3affd01b7e1ee84d", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Candidate 21 source relocation on Python 3.11 with installed dependencies: 23 public imports in two orders and 30 existing API tests. No wheel, dependency isolation or minimum-version claim.", + "title": "Can candidate 21 public APIs import and execute from its relocated source archive without depending on the editable checkout?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "47390b21124f856e", + "event_kind": "finding", + "evidence_refs": [ + "research/package-isolation/REPORT.md", + "research/package-isolation/observations.json", + "research/package-isolation/input-manifest.json", + "research/package-isolation/pytest-stdout.txt", + "research/api-guides/candidate-21-manifest.json" + ], + "execution_observation": { + "candidate_sha256": "e36d69cc5f7d5fe8a063fa7d6289256b2c36bc1413e20840f1d64614068844b4", + "change_reason": "Declared hatchling build backend is unavailable; scoped source isolation replaces only the diagnostic, not wheel acceptance.", + "elapsed_seconds": 6.021768378006527, + "evidence_refs": [ + "research/api-guides/candidate-21-manifest.json", + "research/package-isolation/input-manifest.json", + "research/package-isolation/observations.json" + ], + "executed_method_id": "python.relocated-source-api", + "exit_code": 0, + "input_revision": "sha256:494a303b33222bbc2c2b1ab0ad4bb65879cd46b30ed3dd36312a9024195be7c1", + "limit_seconds": null, + "planned_method_id": "python.offline-package-and-source-check", + "result_candidate_sha256": "e36d69cc5f7d5fe8a063fa7d6289256b2c36bc1413e20840f1d64614068844b4", + "result_ref": "research/package-isolation/observations.json", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Source-isolation checks pass; wheel validation unavailable. Candidate unchanged and all six requirements retained." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:e36d69cc5f7d5fe8a063fa7d6289256b2c36bc1413e20840f1d64614068844b4", + "comparison_scope": "local:relocated-source-api:sha256:494a303b33222bbc2c2b1ab0ad4bb65879cd46b30ed3dd36312a9024195be7c1", + "evaluation_ref": "python:3.11:isolated-source-api:v1", + "evidence_ref": "research/package-isolation/observations.json", + "higher_is_better": true, + "input_revision": "sha256:494a303b33222bbc2c2b1ab0ad4bb65879cd46b30ed3dd36312a9024195be7c1", + "metric": "isolated_existing_api_tests_passed", + "unit": "tests", + "value": 30 + } + ], + "node_id": "question_dbc149db90f786f07c47493d05de4541b670ad9e055813f7cf1ef7de8789148e", + "recorded_at": "2026-10-23T21:43:20+00:00", + "result_id": "finding_78224ec637c9d1fb4e9531683e65c30563aecb68fc730679295399b2fdb4f457", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:803bd35c5d6426a5", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:494a303b33222bbc2c2b1ab0ad4bb65879cd46b30ed3dd36312a9024195be7c1 Applicability: Candidate 21 source relocation on Python 3.11 with installed dependencies: 23 public imports in two orders and 30 existing API tests. No wheel, dependency isolation or minimum-version claim. Observation: Both fresh import orders pass for 23 public modules, with all 132 loaded package origins in relocated source. All 30 existing API tests pass; all 220 candidate files remain unchanged. Hatchling is absent, so wheel validation is unavailable. Interpretation: Checkout dependence is not reproduced in the tested paths. Local HTTP/WSGI/ASGI and nested pytest fixture execution pass from the source archive. The declared build backend is unavailable; no download, substitute backend or wheel-build claim is made. This new probe family does not expand implementation coverage merely by reusing 30 tests. The periodic review retains the original objective, all six requirements and the integration Todo. Recent public-contract fixes remain unsubmitted, native best is still the 6/18 tie, and no hidden assertion is inferred. Next reconcile native history and evaluate candidate 21 only when eligible, retaining exact source/result associations. If gaps remain, select a discriminating numbered-contract probe. Package-build uncertainty gates only that validation path.", + "tags": [ + "writeback-result" + ], + "title": "Can candidate 21 public APIs import and execute from its relocated source archive without depending on the editable checkout?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "1f1e159676844371", + "event_kind": "node", + "evidence_refs": [ + "research/metric-lifetime/REPORT.md" + ], + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T21:57:42Z", + "result_id": "node_c29ea116683d", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:82ad5b3b38f0cc1c", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Candidate 21 fails seven of eleven public metric-lifetime regressions; candidate 22 passes all eleven and the 690-case focused suite. Scope: local tests and preserved source archives, without native failure attribution.", + "title": "Do metric targets include only successful observations from the current generated scenario?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "1da5408fd9de3aab", + "event_kind": "finding", + "evidence_refs": [ + "research/metric-lifetime/REPORT.md", + "research/metric-lifetime/full-test-results.txt", + "research/metric-lifetime/baseline-provenance.json", + "research/metric-lifetime/evaluation-input-manifest.json" + ], + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:c3b928b2bfdb0a6232cf42171b4e32a87819f0ee3a6e2f2c43db59df49919b8e", + "comparison_scope": "local:focused-roadmap-suite:sha256:54fb443f23263b82cadb974db5d50eb464c5b7b9dadd9c7a89dc9feb8c12a56a", + "evaluation_ref": "pytest:8.4.2:python:3.11:noconftest:plugin-autoload-disabled:roadmap", + "evidence_ref": "research/metric-lifetime/full-test-results.txt", + "higher_is_better": true, + "input_revision": "sha256:54fb443f23263b82cadb974db5d50eb464c5b7b9dadd9c7a89dc9feb8c12a56a", + "metric": "focused_tests_passed", + "unit": "tests", + "value": 690 + } + ], + "node_id": "node_c29ea116683d", + "recorded_at": "2026-10-23T21:58:02Z", + "result_id": "finding_d13bad4fbe96", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:ef2dc242dee7d792", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "On preserved candidate 21, four three-measurement scenarios target 3, 6, 9 and 12; empty collectors target synthetic zeros. Seven of eleven regressions fail. Candidate 22 passes all eleven and all 690 focused tests. Real zeros, callback errors and iterator selection are covered. This local suite has its own input revision and proves no native improvement.", + "title": "Candidate 22 resets metric observations per scenario and omits unobserved metrics" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "ac1e6f8ff4755bbc", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T21:58:21+00:00", + "result_id": "question_6a47f5fa907accc4b6b8bcb4f61b1af653897bd648ef2bedb870516dc8ad3643", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:d032fbd36017dc2b", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Public native report 949a49b8e1f7 matched to candidate-21.zip. Phase totals only, no individual assertion attribution; current candidate 22 has separate local evidence.", + "title": "Does preserved candidate 21 satisfy all six roadmap phases in native submission 949a49b8e1f7?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "26931e5d6ae13d80", + "event_kind": "finding", + "evidence_refs": [ + "research/native-candidate-21/REPORT.md", + "research/native-candidate-21/native-result-first.txt", + "research/native-candidate-21/submission-source.json", + "research/native-candidate-21/native-summary.json", + "research/native-candidate-21/source-comparison.json", + "research/metric-lifetime/REPORT.md" + ], + "execution_observation": { + "candidate_sha256": "e36d69cc5f7d5fe8a063fa7d6289256b2c36bc1413e20840f1d64614068844b4", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "research/native-candidate-21/REPORT.md", + "research/native-candidate-21/native-result-first.txt", + "research/native-candidate-21/submission-source.json" + ], + "executed_method_id": "sforge-submit.native", + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "limit_seconds": null, + "planned_method_id": "sforge-submit.native", + "result_candidate_sha256": "e36d69cc5f7d5fe8a063fa7d6289256b2c36bc1413e20840f1d64614068844b4", + "result_ref": "research/native-candidate-21/native-result-first.txt", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Accepted native evaluation completed with a best-score tie and five incomplete phases." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:e36d69cc5f7d5fe8a063fa7d6289256b2c36bc1413e20840f1d64614068844b4", + "comparison_scope": "native:949a49b8e1f7:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-21/native-result-first.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "weighted_score", + "unit": "ratio", + "value": 0.786257 + }, + { + "candidate_ref": "sha256:e36d69cc5f7d5fe8a063fa7d6289256b2c36bc1413e20840f1d64614068844b4", + "comparison_scope": "native:949a49b8e1f7:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-21/native-result-first.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "complete_phases", + "unit": "phases", + "value": 1 + }, + { + "candidate_ref": "sha256:e36d69cc5f7d5fe8a063fa7d6289256b2c36bc1413e20840f1d64614068844b4", + "comparison_scope": "native:949a49b8e1f7:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-21/native-result-first.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "configuration_checks_passed", + "unit": "checks", + "value": 19 + }, + { + "candidate_ref": "sha256:e36d69cc5f7d5fe8a063fa7d6289256b2c36bc1413e20840f1d64614068844b4", + "comparison_scope": "native:949a49b8e1f7:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-21/native-result-first.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "api_checks_passed", + "unit": "checks", + "value": 26 + }, + { + "candidate_ref": "sha256:e36d69cc5f7d5fe8a063fa7d6289256b2c36bc1413e20840f1d64614068844b4", + "comparison_scope": "native:949a49b8e1f7:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-21/native-result-first.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "metrics_checks_passed", + "unit": "checks", + "value": 14 + }, + { + "candidate_ref": "sha256:e36d69cc5f7d5fe8a063fa7d6289256b2c36bc1413e20840f1d64614068844b4", + "comparison_scope": "native:949a49b8e1f7:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-21/native-result-first.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "transport_checks_passed", + "unit": "checks", + "value": 14 + }, + { + "candidate_ref": "sha256:e36d69cc5f7d5fe8a063fa7d6289256b2c36bc1413e20840f1d64614068844b4", + "comparison_scope": "native:949a49b8e1f7:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-21/native-result-first.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "response_checks_passed", + "unit": "checks", + "value": 15 + }, + { + "candidate_ref": "sha256:e36d69cc5f7d5fe8a063fa7d6289256b2c36bc1413e20840f1d64614068844b4", + "comparison_scope": "native:949a49b8e1f7:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-21/native-result-first.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "checks_checks_passed", + "unit": "checks", + "value": 16 + } + ], + "node_id": "question_6a47f5fa907accc4b6b8bcb4f61b1af653897bd648ef2bedb870516dc8ad3643", + "recorded_at": "2026-10-23T21:58:21+00:00", + "result_id": "finding_7c529633219a503191dacd6a1b56c2d32bec6dd2efb83c830555b543988e94a1", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:1c215e7ee920365d", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89 Applicability: Public native report 949a49b8e1f7 matched to candidate-21.zip. Phase totals only, no individual assertion attribution; current candidate 22 has separate local evidence. Observation: Candidate 21 scores 0.786257 with 1/6 complete phases: configuration 19/30, API 26/26, metrics 14/15, transport 14/19, Response 15/18, checks 16/18. Every public count ties native best candidates 6 and 18. Interpretation: The scoped completion question has a negative answer. All 220 captured files matched before and after the single accepted submission; one exact-ID lookup returned its completed result. Retain candidates 6, 18 and 21 as tied native bests. Unknown evaluator revision means separate measurement scopes; no patch is credited with closing an undisclosed assertion. Independent work after this result reproduced synthetic metric targets and stateful scenario accumulation. Candidate 22 corrects those defects and passes 690 focused tests, but has no native result. Its separate source/input identities and baseline comparison are preserved in research/metric-lifetime. Local coverage growth does not establish native progress. Continue public-contract review during cooldown and evaluate the retained current candidate when eligible; keep all six requirements and the integration Todo open.", + "tags": [ + "writeback-result" + ], + "title": "Does preserved candidate 21 satisfy all six roadmap phases in native submission 949a49b8e1f7?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "86fedc2f52763d68", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T22:07:16+00:00", + "result_id": "question_353a2b874cab2a4893368e6111d558af782ebe23db7a7d1c85834fb5011cd0c7", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:2bc6eaa20bb09e3d", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Candidate 22, two migrated repository modules and isolated original fixtures: 108 requests/parser cases, then the same 96 response cases with unified wrappers. No full legacy-suite claim.", + "title": "Do existing repository response-check and content-type regressions preserve their results through v4 namespaces and unified Response objects?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "296244f4d95d0e58", + "event_kind": "finding", + "evidence_refs": [ + "research/legacy-check-integration/REPORT.md", + "research/legacy-check-integration/observations.json", + "research/legacy-check-integration/composite-input-manifest.json", + "research/legacy-check-integration/original-manifest.json", + "research/legacy-check-integration/test-results.txt", + "research/legacy-check-integration/unified-test-results.txt", + "research/metric-lifetime/candidate-22-manifest.json" + ], + "execution_observation": { + "candidate_sha256": "c3b928b2bfdb0a6232cf42171b4e32a87819f0ee3a6e2f2c43db59df49919b8e", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "research/legacy-check-integration/composite-input-manifest.json", + "research/legacy-check-integration/observations.json", + "research/metric-lifetime/candidate-22-manifest.json" + ], + "executed_method_id": "pytest.migrated-repository-checks", + "exit_code": 0, + "input_revision": "sha256:70006188ea745a42da4d6b95a517be3eab536a171aaaff9c8f59ef6950f7f479", + "limit_seconds": null, + "planned_method_id": "pytest.migrated-repository-checks", + "result_candidate_sha256": "c3b928b2bfdb0a6232cf42171b4e32a87819f0ee3a6e2f2c43db59df49919b8e", + "result_ref": "research/legacy-check-integration/observations.json", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Both local runs pass; separate subprocess timings are preserved. Candidate unchanged; native cooldown active." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:c3b928b2bfdb0a6232cf42171b4e32a87819f0ee3a6e2f2c43db59df49919b8e", + "comparison_scope": "local:migrated-repository-checks:sha256:70006188ea745a42da4d6b95a517be3eab536a171aaaff9c8f59ef6950f7f479", + "evaluation_ref": "pytest:8.4.2:python:3.11:isolated-original-fixtures:v1", + "evidence_ref": "research/legacy-check-integration/observations.json", + "higher_is_better": true, + "input_revision": "sha256:70006188ea745a42da4d6b95a517be3eab536a171aaaff9c8f59ef6950f7f479", + "metric": "requests_and_parser_regressions_passed", + "unit": "tests", + "value": 108 + }, + { + "candidate_ref": "sha256:c3b928b2bfdb0a6232cf42171b4e32a87819f0ee3a6e2f2c43db59df49919b8e", + "comparison_scope": "local:migrated-repository-checks:sha256:70006188ea745a42da4d6b95a517be3eab536a171aaaff9c8f59ef6950f7f479", + "evaluation_ref": "pytest:8.4.2:python:3.11:isolated-original-fixtures:v1", + "evidence_ref": "research/legacy-check-integration/observations.json", + "higher_is_better": true, + "input_revision": "sha256:70006188ea745a42da4d6b95a517be3eab536a171aaaff9c8f59ef6950f7f479", + "metric": "unified_response_regressions_passed", + "unit": "tests", + "value": 96 + } + ], + "node_id": "question_353a2b874cab2a4893368e6111d558af782ebe23db7a7d1c85834fb5011cd0c7", + "recorded_at": "2026-10-23T22:07:16+00:00", + "result_id": "finding_6684f904bf16862f9ad18073f97ce4091640861d61ccbd62a3213f55b3cdb350", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:9574d6703a8f24be", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:70006188ea745a42da4d6b95a517be3eab536a171aaaff9c8f59ef6950f7f479 Applicability: Candidate 22, two migrated repository modules and isolated original fixtures: 108 requests/parser cases, then the same 96 response cases with unified wrappers. No full legacy-suite claim. Observation: All 108 migrated repository cases pass; all 96 response cases also pass using unified Responses. Original test functions, parameters and assertions are retained, with one documented header-access adjustment. Candidate 22 is unchanged. Interpretation: The scoped legacy regression hypothesis is not reproduced. Mechanical loader/enum/helper path changes and isolated fixture extraction permit execution without restoring removed APIs. Both positive and negative validation cases cover status, content type, schemas, external references, headers and deduplication. Unified wrappers use the documented lowercase list-valued header access in one diagnostic assertion. Original files and exact patches are preserved; no test case was removed. Two marker and four deprecation warnings remain per run. The 96 cases repeat existing expectations with another representation, not new independent requirements. These local input sets form an explicit composite; no comparison to native scores or earlier test totals is inferred. Full legacy collection remains unavailable, and the 690-case suite was not rerun on unchanged source. Candidate 22 is unsubmitted. Continue with remaining serialization regressions during cooldown while retaining all six requirements and the open Todo.", + "tags": [ + "writeback-result" + ], + "title": "Do existing repository response-check and content-type regressions preserve their results through v4 namespaces and unified Response objects?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "ace3c599fa2b7538", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T22:17:13+00:00", + "result_id": "question_c54df13954ac37ec10bacfc7aef99e706d2b9da77de68641d10b32a3b3025d18", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:9a998aad4b48f690", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Candidate 22: 13 standalone payload functions and two adapted CSV functions, tested across local HTTP/WSGI/ASGI and two OpenAPI versions. Excludes CLI snapshots and the removed class protocol.", + "title": "Do existing repository payload and CSV regressions preserve their expectations through v4 transports, function serializers and unified Responses?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "2a8c3a2e622be6fb", + "event_kind": "finding", + "evidence_refs": [ + "research/legacy-serializer-integration/REPORT.md", + "research/legacy-serializer-integration/observations.json", + "research/legacy-serializer-integration/composite-input-manifest.json", + "research/legacy-serializer-integration/test-results.txt", + "research/legacy-serializer-integration/csv-test-results.txt", + "research/legacy-serializer-integration/initial-proxy-run/test-results.txt", + "research/metric-lifetime/candidate-22-manifest.json" + ], + "execution_observation": { + "candidate_sha256": "c3b928b2bfdb0a6232cf42171b4e32a87819f0ee3a6e2f2c43db59df49919b8e", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "research/legacy-serializer-integration/composite-input-manifest.json", + "research/legacy-serializer-integration/observations.json", + "research/metric-lifetime/candidate-22-manifest.json" + ], + "executed_method_id": "pytest.migrated-repository-serialization", + "exit_code": 0, + "input_revision": "sha256:e442a7b90cf792193cb12cc2365f2ff20f97f29f05351d3c0ba458f3576bd9e0", + "limit_seconds": null, + "planned_method_id": "pytest.migrated-repository-serialization", + "result_candidate_sha256": "c3b928b2bfdb0a6232cf42171b4e32a87819f0ee3a6e2f2c43db59df49919b8e", + "result_ref": "research/legacy-serializer-integration/observations.json", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Both migrated regression runs pass; separate timings and initial environment failures preserved." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:c3b928b2bfdb0a6232cf42171b4e32a87819f0ee3a6e2f2c43db59df49919b8e", + "comparison_scope": "local:migrated-repository-serialization:sha256:e442a7b90cf792193cb12cc2365f2ff20f97f29f05351d3c0ba458f3576bd9e0", + "evaluation_ref": "pytest:8.4.2:python:3.11:local-proxy-bypass:v1", + "evidence_ref": "research/legacy-serializer-integration/observations.json", + "higher_is_better": true, + "input_revision": "sha256:e442a7b90cf792193cb12cc2365f2ff20f97f29f05351d3c0ba458f3576bd9e0", + "metric": "payload_regressions_passed", + "unit": "tests", + "value": 42 + }, + { + "candidate_ref": "sha256:c3b928b2bfdb0a6232cf42171b4e32a87819f0ee3a6e2f2c43db59df49919b8e", + "comparison_scope": "local:migrated-repository-serialization:sha256:e442a7b90cf792193cb12cc2365f2ff20f97f29f05351d3c0ba458f3576bd9e0", + "evaluation_ref": "pytest:8.4.2:python:3.11:local-proxy-bypass:v1", + "evidence_ref": "research/legacy-serializer-integration/observations.json", + "higher_is_better": true, + "input_revision": "sha256:e442a7b90cf792193cb12cc2365f2ff20f97f29f05351d3c0ba458f3576bd9e0", + "metric": "csv_integration_cases_passed", + "unit": "tests", + "value": 12 + } + ], + "node_id": "question_c54df13954ac37ec10bacfc7aef99e706d2b9da77de68641d10b32a3b3025d18", + "recorded_at": "2026-10-23T22:17:13+00:00", + "result_id": "finding_c30e28b6721f3029e2c5f9e11663667fdb97b996bb39a3160560711f4ef5b650", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:a634dde4dfc9bdda", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:e442a7b90cf792193cb12cc2365f2ff20f97f29f05351d3c0ba458f3576bd9e0 Applicability: Candidate 22: 13 standalone payload functions and two adapted CSV functions, tested across local HTTP/WSGI/ASGI and two OpenAPI versions. Excludes CLI snapshots and the removed class protocol. Observation: All 42 payload and 12 CSV cases pass on unchanged candidate 22. Four initial failures were caused by proxy responses to the localhost refusal helper; localhost bypass preserves its original expected exception. All originals and failed inputs are retained. Interpretation: The scoped regression concern is not reproduced. Payload tests retain assertions after namespace and registry-access changes. CSV retains its equality and missing-serializer assertions, while its obsolete class callback becomes a bytes function and Response access uses json(). Original schema generation is copied verbatim; minimal local app fixtures are explicit adaptations. Positive cases exercise all three transports and both schema versions; subsequent negative cases check registry cleanup. The proxy diagnostic and initial 38-pass/four-fail run are preserved with exact input bytes. Final runs each have three deprecation warnings, no skips or xfails. Timings remain separate. No native comparison, hidden-test inference or full-suite completion follows from these local counts. Candidate 22 remains unsubmitted; native 21 remains tied best. Next inspect remaining model/stateful integration regressions and evaluate candidate 22 only when eligible.", + "tags": [ + "writeback-result" + ], + "title": "Do existing repository payload and CSV regressions preserve their expectations through v4 transports, function serializers and unified Responses?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "def107fab8ef4621", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T22:25:18+00:00", + "result_id": "question_ccd51a47327b931c2f650c7012b54ad5c901157c3d48a996b838681998d18f72", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:d1681322ba36fdff", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Hypothesis 6.155.2; 61 cases from 34 existing model/stateful functions with original fixtures and namespace migrations, plus the 690-test focused suite. No older dependency or full-suite claim.", + "title": "Do current Hypothesis hooks preserve linked stateful execution, metric callbacks and settings output through the migrated APIs?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "b94a5552d418dbe6", + "event_kind": "finding", + "evidence_refs": [ + "research/legacy-model-stateful/REPORT.md", + "research/legacy-model-stateful/observations.json", + "research/legacy-model-stateful/evaluation-input-manifest.json", + "research/legacy-model-stateful/candidate-23-manifest.json", + "research/legacy-model-stateful/repository-test-results.txt", + "research/legacy-model-stateful/focused-test-results.txt", + "research/legacy-model-stateful/baseline-provenance.json", + "research/legacy-model-stateful/baseline-test-results.txt" + ], + "execution_observation": { + "candidate_sha256": "edcacdf0835fce2a89c138973ce0ec544e1b412c4f27c9c3c5f8a4ddd550e611", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "research/legacy-model-stateful/baseline-provenance.json", + "research/legacy-model-stateful/candidate-23-manifest.json", + "research/legacy-model-stateful/evaluation-input-manifest.json", + "research/legacy-model-stateful/observations.json" + ], + "executed_method_id": "pytest.model-stateful-integration", + "exit_code": 0, + "input_revision": "sha256:b2d8d9c73ebc8c80eee90c78ce6719308b517a470e55b0a52a012ac184d7e4b4", + "limit_seconds": null, + "planned_method_id": "pytest.model-stateful-integration", + "result_candidate_sha256": "edcacdf0835fce2a89c138973ce0ec544e1b412c4f27c9c3c5f8a4ddd550e611", + "result_ref": "research/legacy-model-stateful/observations.json", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Both current suites pass and controlled archived baseline reproduces eight failures; separate timings retained." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:edcacdf0835fce2a89c138973ce0ec544e1b412c4f27c9c3c5f8a4ddd550e611", + "comparison_scope": "local:model-stateful-integration:sha256:b2d8d9c73ebc8c80eee90c78ce6719308b517a470e55b0a52a012ac184d7e4b4", + "evaluation_ref": "pytest:8.4.2:python:3.11:hypothesis:6.155.2:isolated-fixtures:v1", + "evidence_ref": "research/legacy-model-stateful/observations.json", + "higher_is_better": true, + "input_revision": "sha256:b2d8d9c73ebc8c80eee90c78ce6719308b517a470e55b0a52a012ac184d7e4b4", + "metric": "repository_regressions_passed", + "unit": "tests", + "value": 61 + }, + { + "candidate_ref": "sha256:edcacdf0835fce2a89c138973ce0ec544e1b412c4f27c9c3c5f8a4ddd550e611", + "comparison_scope": "local:model-stateful-integration:sha256:b2d8d9c73ebc8c80eee90c78ce6719308b517a470e55b0a52a012ac184d7e4b4", + "evaluation_ref": "pytest:8.4.2:python:3.11:hypothesis:6.155.2:isolated-fixtures:v1", + "evidence_ref": "research/legacy-model-stateful/observations.json", + "higher_is_better": true, + "input_revision": "sha256:b2d8d9c73ebc8c80eee90c78ce6719308b517a470e55b0a52a012ac184d7e4b4", + "metric": "focused_regressions_passed", + "unit": "tests", + "value": 690 + } + ], + "node_id": "question_ccd51a47327b931c2f650c7012b54ad5c901157c3d48a996b838681998d18f72", + "recorded_at": "2026-10-23T22:25:18+00:00", + "result_id": "finding_f767c38aa803e22dfdea3c56a7aeb1888cac25c50f52dd8583298ac726ef8dc9", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:af252ea1f94143a1", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:b2d8d9c73ebc8c80eee90c78ce6719308b517a470e55b0a52a012ac184d7e4b4 Applicability: Hypothesis 6.155.2; 61 cases from 34 existing model/stateful functions with original fixtures and namespace migrations, plus the 690-test focused suite. No older dependency or full-suite claim. Observation: Candidate 22 repeats eight failures and 53 passes with exact migrated inputs. Candidate 23 fixes result routing and health-check container shape: all 61 existing cases and 690 focused tests pass. Two production files change; original tests remain intact. Interpretation: Six stateful failures arise because installed Hypothesis calls its plural result hook, bypassing the existing singular override. Responses populate catch_all, block root rules and never reach linked rules. A plural hook routes each non-None result by response matcher; the old hook remains unchanged. Two settings failures arise from newer tuple normalization; explicit list conversion preserves helper output. Both defects are grounded in existing assertions, not inferred native tests. The 34 selected functions retain full AST nodes after namespace substitutions. Original app/runner fixtures are retained; optional root-plugin dependencies are isolated. Controlled replay verifies archived candidate 22 and source import origin with identical test bytes. Final runs have no skips; timings remain separate and are not speed comparisons. Candidate 23 is retained with 242 verified files but remains unsubmitted. Other model/stateful cases, full legacy collection and older dependency execution remain unverified; continue those scoped integrations and native evaluation when eligible.", + "tags": [ + "writeback-result" + ], + "title": "Do current Hypothesis hooks preserve linked stateful execution, metric callbacks and settings output through the migrated APIs?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "4a03507a0afe88a0", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T22:42:08+00:00", + "result_id": "question_2037d3b7adbc2d224dde7e789cfff40659b35a44543783b705206ccaf5458c92", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:e6fe924ef916f70f", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Candidate 23/24 with identical 156-case integration inputs on Hypothesis 6.155.2; 144 existing cases and 12 WSGI auth probes. Separate 690-test validation; no full-suite or native claim.", + "title": "Can authentication probes update unified response evidence while preserving check context and handling absent case headers?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "28ee25536421de19", + "event_kind": "finding", + "evidence_refs": [ + "research/stateful-failure-integration/REPORT.md", + "research/stateful-failure-integration/observations.json", + "research/stateful-failure-integration/evaluation-input-manifest.json", + "research/stateful-failure-integration/candidate-24-manifest.json", + "research/stateful-failure-integration/repository-test-results.txt", + "research/stateful-failure-integration/focused-test-results.txt", + "research/stateful-failure-integration/baseline-provenance.json", + "research/stateful-failure-integration/baseline-test-results.txt" + ], + "execution_observation": { + "candidate_sha256": "1a7fa688c9c27294e58eaa7f423cb7a254f218235f60c2197bb7aadf90433482", + "change_reason": null, + "elapsed_seconds": 48.82890494000458, + "evidence_refs": [ + "research/stateful-failure-integration/baseline-provenance.json", + "research/stateful-failure-integration/candidate-24-manifest.json", + "research/stateful-failure-integration/evaluation-input-manifest.json", + "research/stateful-failure-integration/observations.json" + ], + "executed_method_id": "pytest.auth-stateful-integration", + "exit_code": 1, + "input_revision": "sha256:b6c1aad29f3c4b7c3f2f8e6e6e19a0d0a7ab359e52c6292194cad58db917c7cb", + "limit_seconds": null, + "planned_method_id": "pytest.auth-stateful-integration", + "result_candidate_sha256": "1a7fa688c9c27294e58eaa7f423cb7a254f218235f60c2197bb7aadf90433482", + "result_ref": "research/stateful-failure-integration/observations.json", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Thirteen reproduced failures resolved; one unchanged health-check compatibility failure retained for follow-up." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:1a7fa688c9c27294e58eaa7f423cb7a254f218235f60c2197bb7aadf90433482", + "comparison_scope": "local:auth-stateful-integration:sha256:b6c1aad29f3c4b7c3f2f8e6e6e19a0d0a7ab359e52c6292194cad58db917c7cb", + "evaluation_ref": "pytest:8.4.2:python:3.11:hypothesis:6.155.2:isolated-fixtures:v1", + "evidence_ref": "research/stateful-failure-integration/observations.json", + "higher_is_better": true, + "input_revision": "sha256:b6c1aad29f3c4b7c3f2f8e6e6e19a0d0a7ab359e52c6292194cad58db917c7cb", + "metric": "integration_cases_passed", + "unit": "tests", + "value": 155 + } + ], + "node_id": "question_2037d3b7adbc2d224dde7e789cfff40659b35a44543783b705206ccaf5458c92", + "recorded_at": "2026-10-23T22:42:08+00:00", + "result_id": "finding_e3c55a9300bd121fd5f8ccea15eb6d5b1ce9af7a661526f84cbf2434b3b0b0df", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:0a54276dbdf3e790", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:b6c1aad29f3c4b7c3f2f8e6e6e19a0d0a7ab359e52c6292194cad58db917c7cb Applicability: Candidate 23/24 with identical 156-case integration inputs on Hypothesis 6.155.2; 144 existing cases and 12 WSGI auth probes. Separate 690-test validation; no full-suite or native claim. Observation: Candidate 24 resolves 13 authentication-related failures: 155 integration cases pass versus 142 on candidate 23. Both retain one health-check failure. All 690 focused tests pass. One production file changes; original assertions remain intact. Interpretation: Unified Response has slots, so the old update via __dict__ crashes. Reinitialization preserves response identity and refreshes evidence caches. Copying transport kwargs prevents header/session removal from the caller context; initializing absent containers avoids a second invalid-auth crash. Twelve real WSGI cases cover missing/invalid/enforced authentication, absent/empty case headers and primed/unprimed caches. The original stateful ignored-auth failure also passes. Controlled replay verifies archived source and import origin with identical test bytes. Twenty-five additional existing functions retain their assertions and parameters after documented namespace migration; original fixtures and retry behavior are preserved. The one remaining health-check failure is not hidden: installed Hypothesis gives identical attributes for omitted and explicitly empty suppression, so established defaults are unchanged. Timings are separate, not speed comparisons. Candidate 24 remains unsubmitted. Continue remaining regressions and native evaluation when eligible.", + "tags": [ + "writeback-result" + ], + "title": "Can authentication probes update unified response evidence while preserving check context and handling absent case headers?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "331b30f6237704f0", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T22:56:44+00:00", + "result_id": "question_c06157b58ec8f1a1e6dedda587abe9923162503dc754fff4c8c4619657d2422d", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:b5e6e19555d82405", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Public native report 67604c806ad1 bound to candidate 24; phase totals only. Candidate 25 is a separate local correction and has no native result.", + "title": "Does preserved candidate 24 satisfy all six roadmap phases in native submission 67604c806ad1?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "a3b1afea3e7dc75d", + "event_kind": "finding", + "evidence_refs": [ + "research/native-candidate-24/REPORT.md", + "research/native-candidate-24/native-result-second.txt", + "research/native-candidate-24/submission-source.json", + "research/native-candidate-24/native-summary.json", + "research/metric-storage-recovery/REPORT.md", + "research/metric-storage-recovery/observations.json", + "research/metric-storage-recovery/evaluation-input-manifest.json", + "research/metric-storage-recovery/candidate-25-manifest.json" + ], + "execution_observation": { + "candidate_sha256": "1a7fa688c9c27294e58eaa7f423cb7a254f218235f60c2197bb7aadf90433482", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "research/native-candidate-24/native-result-second.txt", + "research/native-candidate-24/submission-source.json" + ], + "executed_method_id": "sforge-submit.native", + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "limit_seconds": null, + "planned_method_id": "sforge-submit.native", + "result_candidate_sha256": "1a7fa688c9c27294e58eaa7f423cb7a254f218235f60c2197bb7aadf90433482", + "result_ref": "research/native-candidate-24/native-result-second.txt", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Accepted evaluation completed with a metrics regression and five incomplete phases; locally corrected successor remains unsubmitted." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:1a7fa688c9c27294e58eaa7f423cb7a254f218235f60c2197bb7aadf90433482", + "comparison_scope": "native:67604c806ad1:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-24/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "weighted_score", + "unit": "ratio", + "value": 0.772924 + }, + { + "candidate_ref": "sha256:1a7fa688c9c27294e58eaa7f423cb7a254f218235f60c2197bb7aadf90433482", + "comparison_scope": "native:67604c806ad1:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-24/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "complete_phases", + "unit": "phases", + "value": 1 + }, + { + "candidate_ref": "sha256:1a7fa688c9c27294e58eaa7f423cb7a254f218235f60c2197bb7aadf90433482", + "comparison_scope": "native:67604c806ad1:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-24/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "configuration_checks_passed", + "unit": "checks", + "value": 19 + }, + { + "candidate_ref": "sha256:1a7fa688c9c27294e58eaa7f423cb7a254f218235f60c2197bb7aadf90433482", + "comparison_scope": "native:67604c806ad1:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-24/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "api_checks_passed", + "unit": "checks", + "value": 26 + }, + { + "candidate_ref": "sha256:1a7fa688c9c27294e58eaa7f423cb7a254f218235f60c2197bb7aadf90433482", + "comparison_scope": "native:67604c806ad1:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-24/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "metrics_checks_passed", + "unit": "checks", + "value": 12 + }, + { + "candidate_ref": "sha256:1a7fa688c9c27294e58eaa7f423cb7a254f218235f60c2197bb7aadf90433482", + "comparison_scope": "native:67604c806ad1:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-24/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "transport_checks_passed", + "unit": "checks", + "value": 14 + }, + { + "candidate_ref": "sha256:1a7fa688c9c27294e58eaa7f423cb7a254f218235f60c2197bb7aadf90433482", + "comparison_scope": "native:67604c806ad1:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-24/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "response_checks_passed", + "unit": "checks", + "value": 15 + }, + { + "candidate_ref": "sha256:1a7fa688c9c27294e58eaa7f423cb7a254f218235f60c2197bb7aadf90433482", + "comparison_scope": "native:67604c806ad1:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-24/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "checks_checks_passed", + "unit": "checks", + "value": 16 + } + ], + "node_id": "question_c06157b58ec8f1a1e6dedda587abe9923162503dc754fff4c8c4619657d2422d", + "recorded_at": "2026-10-23T22:56:44+00:00", + "result_id": "finding_3b37115899e07267383346b04a1f82364960fc1d13f742e6938a917e71919b3c", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:90cc7873715f8e1f", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89 Applicability: Public native report 67604c806ad1 bound to candidate 24; phase totals only. Candidate 25 is a separate local correction and has no native result. Observation: Candidate 24 scores 0.772924 with 1/6 phases complete: configuration 19/30, API 26/26, metrics 12/15, transport 14/19, Response 15/18, checks 16/18. Only metrics falls relative to candidate 21; best remains 0.786257. Interpretation: The scoped acceptance question has a negative answer. One accepted submission was polled by its exact ID, retaining pending and completed outputs. All 266 source-snapshot files matched around upload; native grading duration and evaluator revision are undisclosed. Counts motivate metric review but identify no hidden assertion or cause. The separate controlled 55-case comparison exposes a storage compatibility regression: candidate 21 passes 48, candidate 24 passes 53, and corrected candidate 25 passes all 55. Candidate 25 restores per-metric lists while filtering empty targets and retaining scenario isolation. It also passes 690 focused and 171 integration/auth cases, with the same single health-check failure. Those local results are not this native result and do not prove score recovery. Preserve native-best candidates 6/18/21 and all six requirements. Continue contract-level review and wait for server eligibility before another evaluation.", + "tags": [ + "writeback-result" + ], + "title": "Does preserved candidate 24 satisfy all six roadmap phases in native submission 67604c806ad1?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "b6ceca770e60a2c6", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T23:05:56+00:00", + "result_id": "question_52dcb344c1d0fb897eaf7f1d4148d8073e91c7b7c2d585d24574ed6ee98ab8d5", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:849642ba54712a54", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Python 3.11 executing exact owned decorator expressions and package import under a 3.8/3.9 keyword-signature shim. No old-interpreter, dependency, or native compatibility claim.", + "title": "Do candidate 25 owned dataclass decorators preserve the keyword API available throughout the declared Python 3.8+ range?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "22d5bc59c23a6c1b", + "event_kind": "finding", + "evidence_refs": [ + "research/replan-python-compatibility/REPORT.md", + "research/replan-python-compatibility/observations.json", + "research/replan-python-compatibility/probe.py", + "research/replan-python-compatibility/import_probe.py", + "research/replan-python-compatibility/input-manifest.json", + "research/replan-python-compatibility/source-reuse.json", + "research/metric-storage-recovery/candidate-25-manifest.json" + ], + "execution_observation": { + "candidate_sha256": "6db2c586315cda3a2fb5617ee0df70b01d152b58c1b349fad3d8a4913279c68f", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "research/metric-storage-recovery/candidate-25-manifest.json", + "research/replan-python-compatibility/import_probe.py", + "research/replan-python-compatibility/input-manifest.json", + "research/replan-python-compatibility/observations.json", + "research/replan-python-compatibility/probe.py", + "research/replan-python-compatibility/source-reuse.json" + ], + "executed_method_id": "python.owned-dataclass-keyword-probe", + "exit_code": 0, + "input_revision": "sha256:595176ef46ab0cb823177d2887f9ea9906881a468ebee4c01b90b6e7dc6b7d7f", + "limit_seconds": null, + "planned_method_id": "python.owned-dataclass-keyword-probe", + "result_candidate_sha256": "6db2c586315cda3a2fb5617ee0df70b01d152b58c1b349fad3d8a4913279c68f", + "result_ref": "research/replan-python-compatibility/observations.json", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Bounded review isolates four unsupported decorator sites; compatibility repair selected without production edits." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:6db2c586315cda3a2fb5617ee0df70b01d152b58c1b349fad3d8a4913279c68f", + "comparison_scope": "local:declared-python-dataclass-keywords", + "evaluation_ref": "python:3.11:owned-dataclass-keyword-shim:v1", + "evidence_ref": "research/replan-python-compatibility/observations.json", + "higher_is_better": false, + "input_revision": "sha256:595176ef46ab0cb823177d2887f9ea9906881a468ebee4c01b90b6e7dc6b7d7f", + "metric": "unsupported_owned_decorators", + "unit": "definitions", + "value": 4 + } + ], + "node_id": "question_52dcb344c1d0fb897eaf7f1d4148d8073e91c7b7c2d585d24574ed6ee98ab8d5", + "recorded_at": "2026-10-23T23:05:56+00:00", + "result_id": "finding_23a4201ba78ae0084a06325000a700b405b8fc81ec3830ce1090601eb090e7b0", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:c2599829923d8175", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:595176ef46ab0cb823177d2887f9ea9906881a468ebee4c01b90b6e7dc6b7d7f Applicability: Python 3.11 executing exact owned decorator expressions and package import under a 3.8/3.9 keyword-signature shim. No old-interpreter, dependency, or native compatibility claim. Observation: Four decorators accept the installed API but reject the older keyword signature. Package import fails at CheckContext under the owned-source shim. Two explicit-slot dataclass controls pass. All 281 candidate files remain unchanged. Interpretation: The scoped keyword-compatibility question has a negative answer. pyproject.toml and the CI matrix retain Python 3.8/3.9, while MetricContext, MetricCollector, CheckContext and _ClientSerializer request dataclass slots, introduced in 3.10. The import probe constrains only repository decorator calls; installed dependencies retain their own API to avoid false attribution. An explicit-slot dataclass control preserves field identity and lack of an instance dictionary under both APIs. This isolates one incompatibility, not the complete behavior of absent older interpreters. Reuse the 12 unchanged configuration files behind prior audits and candidate 25 validation; no broad rerun is needed for this unchanged source. Select a four-site repair preserving signatures, defaults, mutable isolation and existing dataclass behavior, then validate imports and behavior. Do not raise the minimum version or infer native score recovery. All six requirements remain open.", + "tags": [ + "writeback-result" + ], + "title": "Do candidate 25 owned dataclass decorators preserve the keyword API available throughout the declared Python 3.8+ range?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "34a44e854a88f205", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T23:13:44+00:00", + "result_id": "question_ef99cb95975bdf2109803465d2b1b884fb2d2756cd11ddb1e6faa1055136170a", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:00bff6ed6b8a04ef", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Archived candidates 25/26 on Python 3.11, identical 24-case inputs; shim restricts owned dataclass keyword calls only. No actual old-interpreter, dependency or native acceptance claim.", + "title": "Does the four-site slots repair preserve dataclass behavior while allowing package import under the Python 3.8/3.9 keyword contract?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "402fc78d10890605", + "event_kind": "finding", + "evidence_refs": [ + "research/python-dataclass-compatibility/REPORT.md", + "research/python-dataclass-compatibility/observations.json", + "research/python-dataclass-compatibility/test_dataclass_contract.py", + "research/python-dataclass-compatibility/keyword_runner.py", + "research/python-dataclass-compatibility/evaluation-input-manifest.json", + "research/python-dataclass-compatibility/candidate-26-manifest.json" + ], + "execution_observation": { + "candidate_sha256": "ccc32023dadee4972484b1e26c6b8754d8e0bc7189936afb4f935b949265cdd6", + "change_reason": null, + "elapsed_seconds": 0.7158851249987492, + "evidence_refs": [ + "research/python-dataclass-compatibility/candidate-26-manifest.json", + "research/python-dataclass-compatibility/evaluation-input-manifest.json", + "research/python-dataclass-compatibility/keyword_runner.py", + "research/python-dataclass-compatibility/observations.json", + "research/python-dataclass-compatibility/test_dataclass_contract.py" + ], + "executed_method_id": "pytest.owned-dataclass-keyword-shim", + "exit_code": 0, + "input_revision": "sha256:78700887bf6906615cba9fda03917b40415775aacc2c47a0b340f0feedc08b8d", + "limit_seconds": null, + "planned_method_id": "pytest.owned-dataclass-keyword-shim", + "result_candidate_sha256": "ccc32023dadee4972484b1e26c6b8754d8e0bc7189936afb4f935b949265cdd6", + "result_ref": "research/python-dataclass-compatibility/observations.json", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Four-site repair passes the fixed behavior probe; broader acceptance and native evaluation remain open." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:ccc32023dadee4972484b1e26c6b8754d8e0bc7189936afb4f935b949265cdd6", + "comparison_scope": "local:dataclass-behavior:owned-keyword-shim", + "evaluation_ref": "python:3.11:pytest:8.4.2:owned-keyword-shim:v1", + "evidence_ref": "research/python-dataclass-compatibility/observations.json", + "higher_is_better": true, + "input_revision": "sha256:78700887bf6906615cba9fda03917b40415775aacc2c47a0b340f0feedc08b8d", + "metric": "behavior_cases_passed_under_keyword_shim", + "unit": "tests", + "value": 24 + } + ], + "node_id": "question_ef99cb95975bdf2109803465d2b1b884fb2d2756cd11ddb1e6faa1055136170a", + "recorded_at": "2026-10-23T23:13:44+00:00", + "result_id": "finding_dd010c447e0c244f8d0e422e7c82929de1d83829c058d2a04f5d767be7f1d709", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:1214f59246c41123", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:78700887bf6906615cba9fda03917b40415775aacc2c47a0b340f0feedc08b8d Applicability: Archived candidates 25/26 on Python 3.11, identical 24-case inputs; shim restricts owned dataclass keyword calls only. No actual old-interpreter, dependency or native acceptance claim. Observation: Candidate 25 fails import under the old keyword signature; candidate 26 imports and passes 24 behavior cases. Both pass those cases normally. Candidate 26 also retains 690 focused and 171 integration/auth passes, with one known health-check failure. Interpretation: The scoped repair succeeds. A slots helper processes ordinary dataclasses after their methods and defaults are generated, preserving the four mutable callers without requiring the Python-3.10 keyword. Identical tests cover signatures, field metadata, no instance dictionary, default isolation, copies, pickle round trips, replacement and serializer execution. Archives and import origins are verified. The first draft isolation assertion overlooked lazy check-config population; its failing input is preserved and the corrected assertion passes on candidate 25 before the repair. Existing assertions are unchanged. This does not certify absent older interpreters or recover a native score. Keep all six roadmap targets open, retain the known health-check failure and best native snapshots, and evaluate the exact preserved successor only when eligible.", + "tags": [ + "writeback-result" + ], + "title": "Does the four-site slots repair preserve dataclass behavior while allowing package import under the Python 3.8/3.9 keyword contract?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "79c0c6116027705e", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T23:22:17+00:00", + "result_id": "question_ab4719a3c261b14d9c35a4cafe30b38847010abc0a33dd0ca7784813cf36adbd", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:785413d07c29cd31", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Archived candidates 26/27 on Python 3.11; identical 55-case reporting inputs cover native and normalized responses, direct checks and enclosing results. No native acceptance claim.", + "title": "Can check serialization preserve explicit request evidence for requestless responses without changing request precedence or missing-evidence errors?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "d37bc78cfa7601e4", + "event_kind": "finding", + "evidence_refs": [ + "research/response-reporting/REPORT.md", + "research/response-reporting/observations.json", + "research/response-reporting/test_request_evidence.py", + "research/response-reporting/migration.json", + "research/response-reporting/evaluation-input-manifest.json", + "research/response-reporting/candidate-27-manifest.json" + ], + "execution_observation": { + "candidate_sha256": "208f7a14e523d42d06dd23e7ef31030c0fdad783d32a5594f458583175da7ed5", + "change_reason": null, + "elapsed_seconds": 1.2219395839929348, + "evidence_refs": [ + "research/response-reporting/candidate-27-manifest.json", + "research/response-reporting/evaluation-input-manifest.json", + "research/response-reporting/migration.json", + "research/response-reporting/observations.json", + "research/response-reporting/test_request_evidence.py" + ], + "executed_method_id": "pytest.archived-reporting", + "exit_code": 0, + "input_revision": "sha256:435374f3189a836ae9d12f2176a4c68a9ba1874a6e8195b53f2a655a84bcddd6", + "limit_seconds": null, + "planned_method_id": "pytest.archived-reporting", + "result_candidate_sha256": "208f7a14e523d42d06dd23e7ef31030c0fdad783d32a5594f458583175da7ed5", + "result_ref": "research/response-reporting/observations.json", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Reporting repair passes the fixed comparison; broader acceptance and native evaluation remain open." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:208f7a14e523d42d06dd23e7ef31030c0fdad783d32a5594f458583175da7ed5", + "comparison_scope": "local:reporting:request-evidence", + "evaluation_ref": "python:3.11:pytest:8.4.2:archived-reporting:v1", + "evidence_ref": "research/response-reporting/observations.json", + "higher_is_better": true, + "input_revision": "sha256:435374f3189a836ae9d12f2176a4c68a9ba1874a6e8195b53f2a655a84bcddd6", + "metric": "reporting_cases_passed", + "unit": "tests", + "value": 55 + } + ], + "node_id": "question_ab4719a3c261b14d9c35a4cafe30b38847010abc0a33dd0ca7784813cf36adbd", + "recorded_at": "2026-10-23T23:22:17+00:00", + "result_id": "finding_b4c573d86affb83b3756d05cc1b77d4367b331179ca4f10e419c37d3ed697e20", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:39214251f463a1b5", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:435374f3189a836ae9d12f2176a4c68a9ba1874a6e8195b53f2a655a84bcddd6 Applicability: Archived candidates 26/27 on Python 3.11; identical 55-case reporting inputs cover native and normalized responses, direct checks and enclosing results. No native acceptance claim. Observation: Candidate 26 passes 23 and fails 32 reporting cases; candidate 27 passes all 55. The latter retains 690 focused and 171 other integration/auth passes, with the known health-check failure unchanged. Interpretation: The serializer now selects the response-attached request first and falls back to the explicit check request only when it is absent. With neither source, the established InternalError is retained; timeout behavior is unchanged. The fixed matrix covers four response sources, status 200/500, request precedence, missing evidence, direct/outer serialization and sanitization without mutation. Five existing repository cases retain their assertions with only namespace migrations. Archived sources, import origins and identical inputs are verified. The initial probe had two setup/expectation mistakes; its bytes and failures are preserved, and the corrected probe reproduced 32 failures on the baseline before the repair. No request is fabricated and history traversal is unchanged. Candidate 27 remains unsubmitted; keep all six targets and the known health-check failure open, then evaluate a preserved successor only when eligible.", + "tags": [ + "writeback-result" + ], + "title": "Can check serialization preserve explicit request evidence for requestless responses without changing request precedence or missing-evidence errors?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "38d139622856662e", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T23:28:49+00:00", + "result_id": "question_3990c9ff1e80c15d6881c9fc89c685c0d5c25f57ba6d24d63fb73867711efa26", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:90e39b5ae6a3dde8", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Unchanged candidate 27 on Python 3.11; nine original test functions with documented local fixture adaptations, plus eight WSGI wire checks. No full-suite or native acceptance claim.", + "title": "Do the previously omitted model network and generated-pytest regressions preserve their assertions through the v4 transport and response integration?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "7a98763314863c4c", + "event_kind": "finding", + "evidence_refs": [ + "research/model-transport-integration/REPORT.md", + "research/model-transport-integration/observations.json", + "research/model-transport-integration/composite-input-manifest.json", + "research/model-transport-integration/migration.json", + "research/model-transport-integration/generated-migration.json", + "research/model-transport-integration/test-results.txt", + "research/model-transport-integration/generated-results.txt" + ], + "execution_observation": { + "candidate_sha256": "208f7a14e523d42d06dd23e7ef31030c0fdad783d32a5594f458583175da7ed5", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "research/model-transport-integration/composite-input-manifest.json", + "research/model-transport-integration/generated-migration.json", + "research/model-transport-integration/generated-results.txt", + "research/model-transport-integration/migration.json", + "research/model-transport-integration/observations.json", + "research/model-transport-integration/test-results.txt" + ], + "executed_method_id": "pytest.model-network-and-generated", + "exit_code": 0, + "input_revision": "sha256:db8085ffe5932bcccad8a2b097a4758d3d95125197ad4506e40be653243f972f", + "limit_seconds": null, + "planned_method_id": "pytest.model-network-and-generated", + "result_candidate_sha256": "208f7a14e523d42d06dd23e7ef31030c0fdad783d32a5594f458583175da7ed5", + "result_ref": "research/model-transport-integration/observations.json", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Both scoped regression runs pass; retain source and continue the open native/full-suite frontier." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:208f7a14e523d42d06dd23e7ef31030c0fdad783d32a5594f458583175da7ed5", + "comparison_scope": "local:model-network-and-generated-pytest", + "evaluation_ref": "python:3.11:pytest:8.4.2:local-fixtures:v1", + "evidence_ref": "research/model-transport-integration/observations.json", + "higher_is_better": true, + "input_revision": "sha256:db8085ffe5932bcccad8a2b097a4758d3d95125197ad4506e40be653243f972f", + "metric": "integration_cases_passed", + "unit": "tests", + "value": 26 + } + ], + "node_id": "question_3990c9ff1e80c15d6881c9fc89c685c0d5c25f57ba6d24d63fb73867711efa26", + "recorded_at": "2026-10-23T23:28:49+00:00", + "result_id": "finding_37142afce43b11aa646aa5b2d5d020c57f936e28cbee7e2f59b452c3a3aa78fa", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:ff8cdeafe8c4df83", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:db8085ffe5932bcccad8a2b097a4758d3d95125197ad4506e40be653243f972f Applicability: Unchanged candidate 27 on Python 3.11; nine original test functions with documented local fixture adaptations, plus eight WSGI wire checks. No full-suite or native acceptance claim. Observation: All 26 cases pass: 18 existing model/generated-pytest cases and eight WSGI override cases. Nine original test functions and their helper match after declared namespace changes. All 296 candidate files remain unchanged. Interpretation: The selected regressions do not reproduce a new migration defect, so retain candidate 27. Original assertions exercise HTTP URLs, headers, cookies, body updates, validation options, missing-client guidance, repeated response headers and exact generated-test failure output. Added WSGI cases check real application inputs and prepared requests for both schema versions and client ownership paths, without relying on obsolete client keyword mocks. Local server, minimal schema and pytester fixture adaptations are explicit; original files and the YAML fixture are preserved. Separate input manifests and process results bind both passing runs to the unchanged archive. This extends the earlier omitted integration slice rather than repeating its tests. Prior focused/reporting evidence and the known health-check failure remain attributed to their original runs. Keep all six requirements open and reconcile native eligibility before evaluating the preserved successor; these local passes do not prove native recovery.", + "tags": [ + "writeback-result" + ], + "title": "Do the previously omitted model network and generated-pytest regressions preserve their assertions through the v4 transport and response integration?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "fd69693aeb9cdbc4", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T23:36:27+00:00", + "result_id": "question_a3d06504765d11ff5af122c4f13f0bafc23cc9cba370da1ac75248b07ee8badd", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:ece87fb258c7534e", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Candidate 27 on Python 3.11 with the original Flask GraphQL app and documented context/fixture adaptations; 25 existing test functions. Excludes CLI snapshots and full-suite/native acceptance.", + "title": "Do existing GraphQL live-app, scalar and generated-pytest regressions retain their assertions through the v4 transport, response and check integration?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "faf22b56f9c7cb42", + "event_kind": "finding", + "evidence_refs": [ + "research/graphql-live-integration/REPORT.md", + "research/graphql-live-integration/observations.json", + "research/graphql-live-integration/input-manifest.json", + "research/graphql-live-integration/migration.json", + "research/graphql-live-integration/test-results.txt" + ], + "execution_observation": { + "candidate_sha256": "208f7a14e523d42d06dd23e7ef31030c0fdad783d32a5594f458583175da7ed5", + "change_reason": null, + "elapsed_seconds": 13.259880090001388, + "evidence_refs": [ + "research/graphql-live-integration/input-manifest.json", + "research/graphql-live-integration/migration.json", + "research/graphql-live-integration/observations.json", + "research/graphql-live-integration/test-results.txt" + ], + "executed_method_id": "pytest.graphql-live-and-generated", + "exit_code": 0, + "input_revision": "sha256:6d90dda9a07f05ed7b79fc95fe9e89d76bdfc1bcd2909038d2ead990d3518b02", + "limit_seconds": null, + "planned_method_id": "pytest.graphql-live-and-generated", + "result_candidate_sha256": "208f7a14e523d42d06dd23e7ef31030c0fdad783d32a5594f458583175da7ed5", + "result_ref": "research/graphql-live-integration/observations.json", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Scoped regressions pass; retain candidate and continue open native/full-suite validation." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:208f7a14e523d42d06dd23e7ef31030c0fdad783d32a5594f458583175da7ed5", + "comparison_scope": "local:graphql-live-scalar-and-generated-pytest", + "evaluation_ref": "python:3.11:pytest:8.4.2:original-graphql-app:v1", + "evidence_ref": "research/graphql-live-integration/observations.json", + "higher_is_better": true, + "input_revision": "sha256:6d90dda9a07f05ed7b79fc95fe9e89d76bdfc1bcd2909038d2ead990d3518b02", + "metric": "integration_cases_passed", + "unit": "tests", + "value": 37 + } + ], + "node_id": "question_a3d06504765d11ff5af122c4f13f0bafc23cc9cba370da1ac75248b07ee8badd", + "recorded_at": "2026-10-23T23:36:27+00:00", + "result_id": "finding_4987159a7d76667945f8af80343b12726ce1f06dd8aba030c333df16d2553e64", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:dab49dc133f5f403", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:6d90dda9a07f05ed7b79fc95fe9e89d76bdfc1bcd2909038d2ead990d3518b02 Applicability: Candidate 27 on Python 3.11 with the original Flask GraphQL app and documented context/fixture adaptations; 25 existing test functions. Excludes CLI snapshots and full-suite/native acceptance. Observation: All 37 selected cases pass. Two nested pytest runs each assert four operation tests and 40 Hypothesis calls. Twenty-five original test functions, three helpers/fixtures, eight original files and all 296 candidate files verify against preserved sources. Interpretation: No new migration defect is reproduced, so retain candidate 27. Coverage includes real query/mutation requests, HTTP/WSGI generated pytest integration, scalar registration and invalid inputs, request bodies, base URLs, introspection failures, missing roots, field lookup and GraphQL response/error checks. Original assertions and the application are preserved; the standalone check receives its required CheckContext, imports point to the copied fixture package, and owned local servers replace background fixture startup. The deprecated-method test and CLI snapshots remain excluded explicitly. Quantitative evidence refers only to this unchanged candidate and exact input manifest; nested generated tests are not double-counted. Prior coverage and the unresolved health-check limitation remain attributed to their original runs. Reconcile server eligibility and evaluate the preserved successor before inferring native recovery; all six requirements remain open.", + "tags": [ + "writeback-result" + ], + "title": "Do existing GraphQL live-app, scalar and generated-pytest regressions retain their assertions through the v4 transport, response and check integration?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "963639f903e1e72b", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T23:47:42+00:00", + "result_id": "question_ffe8858551676b54704bf900b254c9b16c66709d0b18d11782977354981f5d10", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:3de94bf45ab25295", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Archived candidates 27/28 on identical 22-case CLI inputs in Python 3.11, with local HTTP/WSGI/ASGI apps and report validation. Native acceptance is outside this comparison.", + "title": "Do CLI generation modes survive the shared configuration handoff while migrated serializers, checks, metrics and response reports execute across all three transports?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "35ca72c47c70e7be", + "event_kind": "finding", + "evidence_refs": [ + "research/cli-transport-integration/REPORT.md", + "research/cli-transport-integration/observations.json", + "research/cli-transport-integration/evaluation-input-manifest.json", + "research/cli-transport-integration/candidate-28-manifest.json", + "research/cli-transport-integration/candidate-diff.patch", + "research/cli-transport-integration/test_cli_matrix.py", + "research/cli-transport-integration/native-source-comparison.json" + ], + "execution_observation": { + "candidate_sha256": "ea62d085e2976c0b495ef719e9214110c696bd7fc412f2c7bdc17fe911d5e5d5", + "change_reason": null, + "elapsed_seconds": 3.7195082520047436, + "evidence_refs": [ + "research/cli-transport-integration/candidate-28-manifest.json", + "research/cli-transport-integration/candidate-diff.patch", + "research/cli-transport-integration/evaluation-input-manifest.json", + "research/cli-transport-integration/native-source-comparison.json", + "research/cli-transport-integration/observations.json", + "research/cli-transport-integration/test_cli_matrix.py" + ], + "executed_method_id": "pytest.archived-cli-matrix", + "exit_code": 0, + "input_revision": "sha256:924d0b0c0b43d5393f5d52ce272524d0d76f0f8a82392f2bfc93d0eb75d9accc", + "limit_seconds": null, + "planned_method_id": "pytest.archived-cli-matrix", + "result_candidate_sha256": "ea62d085e2976c0b495ef719e9214110c696bd7fc412f2c7bdc17fe911d5e5d5", + "result_ref": "research/cli-transport-integration/observations.json", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "CLI repair and focused validation pass; preserve candidate for the next eligible native evaluation." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:ea62d085e2976c0b495ef719e9214110c696bd7fc412f2c7bdc17fe911d5e5d5", + "comparison_scope": "local:cli-mode-and-transport-integration", + "evaluation_ref": "python:3.11:pytest:8.4.2:archived-cli-matrix:v1", + "evidence_ref": "research/cli-transport-integration/observations.json", + "higher_is_better": true, + "input_revision": "sha256:924d0b0c0b43d5393f5d52ce272524d0d76f0f8a82392f2bfc93d0eb75d9accc", + "metric": "cli_integration_cases_passed", + "unit": "tests", + "value": 22 + } + ], + "node_id": "question_ffe8858551676b54704bf900b254c9b16c66709d0b18d11782977354981f5d10", + "recorded_at": "2026-10-23T23:47:42+00:00", + "result_id": "finding_02e35761a331df51a99c96e1a52b49384871ce43b572029fa70190e3c2605df0", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:e7754296da8a4ef8", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:924d0b0c0b43d5393f5d52ce272524d0d76f0f8a82392f2bfc93d0eb75d9accc Applicability: Archived candidates 27/28 on identical 22-case CLI inputs in Python 3.11, with local HTTP/WSGI/ASGI apps and report validation. Native acceptance is outside this comparison. Observation: Candidate 27 passes seven and fails 15 CLI cases; candidate 28 passes all 22 and the 690 focused cases. The CLI now copies selected modes into GenerationConfig. All 308 candidate files and identical comparison inputs verify. Interpretation: The CLI loader received the selected mode, but runner cloning reapplied GenerationConfig defaults containing both modes. Initializing that config from the CLI selection preserves default positive, explicit positive/negative and all. The fixed matrix checks real request bodies, serializers returning bytes/None, rich metric/check contexts, repeated response headers, exclusions, successful/failed outcomes and JUnit/cassette output across three transports. Archived import origins and inputs are verified. Initial app-loading and lifespan fixture errors are preserved; the early missing-mode explanation was corrected after explicit controls reproduced the production defect. The corrected baseline failed before the one-line edit and reproduces on archived candidate 27. Candidate 28 remains unsubmitted; native history has six completed submissions and still reports cooldown. Keep all six requirements and the health-check limitation open, then evaluate the preserved successor after eligibility readback.", + "tags": [ + "writeback-result" + ], + "title": "Do CLI generation modes survive the shared configuration handoff while migrated serializers, checks, metrics and response reports execute across all three transports?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "1b98bcab800711f9", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-23T23:57:14+00:00", + "result_id": "question_2a00086950da1bbfb259d70a63d192b18be456bed84f709aaf5ecf5d345fb65a", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:c48393f35b445aee", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Public native report 990ed96f6bbb bound to candidate 28; phase totals only. Earlier native and local runs retain their own source identities and unverified cross-run evaluator comparability.", + "title": "Does preserved candidate 28 satisfy all six roadmap phases in native submission 990ed96f6bbb?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "198732a9f308305b", + "event_kind": "finding", + "evidence_refs": [ + "research/native-candidate-28/REPORT.md", + "research/native-candidate-28/native-result-second.txt", + "research/native-candidate-28/submission-source.json", + "research/native-candidate-28/native-summary.json", + "research/native-candidate-28/source-comparison.json", + "research/native-candidate-28/candidate24-to-28.patch", + "research/native-candidate-28/local-evidence-lineage.json" + ], + "execution_observation": { + "candidate_sha256": "ea62d085e2976c0b495ef719e9214110c696bd7fc412f2c7bdc17fe911d5e5d5", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "research/native-candidate-28/native-result-second.txt", + "research/native-candidate-28/submission-source.json" + ], + "executed_method_id": "sforge-submit.native", + "exit_code": 0, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "limit_seconds": null, + "planned_method_id": "sforge-submit.native", + "result_candidate_sha256": "ea62d085e2976c0b495ef719e9214110c696bd7fc412f2c7bdc17fe911d5e5d5", + "result_ref": "research/native-candidate-28/native-result-second.txt", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Accepted evaluation completed at the prior best score; five phases remain incomplete and work continues." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:ea62d085e2976c0b495ef719e9214110c696bd7fc412f2c7bdc17fe911d5e5d5", + "comparison_scope": "native:990ed96f6bbb:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-28/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "weighted_score", + "unit": "ratio", + "value": 0.786257 + }, + { + "candidate_ref": "sha256:ea62d085e2976c0b495ef719e9214110c696bd7fc412f2c7bdc17fe911d5e5d5", + "comparison_scope": "native:990ed96f6bbb:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-28/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "complete_phases", + "unit": "phases", + "value": 1 + }, + { + "candidate_ref": "sha256:ea62d085e2976c0b495ef719e9214110c696bd7fc412f2c7bdc17fe911d5e5d5", + "comparison_scope": "native:990ed96f6bbb:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-28/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "configuration_checks_passed", + "unit": "checks", + "value": 19 + }, + { + "candidate_ref": "sha256:ea62d085e2976c0b495ef719e9214110c696bd7fc412f2c7bdc17fe911d5e5d5", + "comparison_scope": "native:990ed96f6bbb:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-28/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "api_checks_passed", + "unit": "checks", + "value": 26 + }, + { + "candidate_ref": "sha256:ea62d085e2976c0b495ef719e9214110c696bd7fc412f2c7bdc17fe911d5e5d5", + "comparison_scope": "native:990ed96f6bbb:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-28/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "metrics_checks_passed", + "unit": "checks", + "value": 14 + }, + { + "candidate_ref": "sha256:ea62d085e2976c0b495ef719e9214110c696bd7fc412f2c7bdc17fe911d5e5d5", + "comparison_scope": "native:990ed96f6bbb:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-28/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "transport_checks_passed", + "unit": "checks", + "value": 14 + }, + { + "candidate_ref": "sha256:ea62d085e2976c0b495ef719e9214110c696bd7fc412f2c7bdc17fe911d5e5d5", + "comparison_scope": "native:990ed96f6bbb:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-28/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "response_checks_passed", + "unit": "checks", + "value": 15 + }, + { + "candidate_ref": "sha256:ea62d085e2976c0b495ef719e9214110c696bd7fc412f2c7bdc17fe911d5e5d5", + "comparison_scope": "native:990ed96f6bbb:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-28/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "checks_checks_passed", + "unit": "checks", + "value": 16 + } + ], + "node_id": "question_2a00086950da1bbfb259d70a63d192b18be456bed84f709aaf5ecf5d345fb65a", + "recorded_at": "2026-10-23T23:57:14+00:00", + "result_id": "finding_8a2a536eb8c04a401ed666ff55b7f06097da79aa7e7015097e5c6eab1e2a3f1f", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:888b2660e141ef1d", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89 Applicability: Public native report 990ed96f6bbb bound to candidate 28; phase totals only. Earlier native and local runs retain their own source identities and unverified cross-run evaluator comparability. Observation: Candidate 28 scores 0.786257 with one of six phases complete: configuration 19/30, API 26/26, metrics 14/15, transport 14/19, Response 15/18, checks 16/18. It ties the prior best; candidate 24 reported metrics 12/15. Interpretation: The scoped acceptance question has a negative answer: five phases remain incomplete. One accepted submission is recovered by its exact ID; pending and completed outputs remain preserved. All 308 candidate files and 179 Python source files match before upload and after the result. Archives 21/24 remain verified and recoverable. The observed counts support retaining candidate 28 alongside earlier bests, but undisclosed evaluator/input revisions prevent claiming a controlled native delta or attributing recovery to one patch. Source comparison records the metric storage, dataclass compatibility, request evidence and CLI mode changes since 24. Existing local comparisons support those fixes independently and are not rerun or relabeled as native evidence. Keep all original requirements and the health-check limitation open. Next review configuration behavior against prior audits and public regressions, requiring a new reproduction before changes; honor the accepted-submission cooldown.", + "tags": [ + "writeback-result" + ], + "title": "Does preserved candidate 28 satisfy all six roadmap phases in native submission 990ed96f6bbb?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "0b8f9c63d1b76c7d", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-24T00:05:17+00:00", + "result_id": "question_ce0f7b93bf30e98248f7bc3b6d4246adad597b92fdd767f6b84c233192aad56a", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:06be5d2827f0083e", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Candidate 28 on Python 3.11: direct CheckContext object/dictionary construction, with operation/phase resolution and explicit named settings as controls. No native failure attribution.", + "title": "Do direct check-configuration reads preserve the effects of later shared-default edits independently of access order?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "84294b53762901ba", + "event_kind": "finding", + "evidence_refs": [ + "research/replan-check-defaults/REPORT.md", + "research/replan-check-defaults/observations.json", + "research/replan-check-defaults/probe.py", + "research/replan-check-defaults/selection-observations.json", + "research/replan-check-defaults/selection_probe.py", + "research/replan-check-defaults/source-reuse.json", + "research/replan-check-defaults/runtime-provenance.json" + ], + "execution_observation": { + "candidate_sha256": "ea62d085e2976c0b495ef719e9214110c696bd7fc412f2c7bdc17fe911d5e5d5", + "change_reason": null, + "elapsed_seconds": 0.1589220840105554, + "evidence_refs": [ + "research/replan-check-defaults/observations.json", + "research/replan-check-defaults/probe.py", + "research/replan-check-defaults/runtime-provenance.json", + "research/replan-check-defaults/selection-observations.json", + "research/replan-check-defaults/selection_probe.py", + "research/replan-check-defaults/source-reuse.json" + ], + "executed_method_id": "python.check-default-read-order", + "exit_code": 0, + "input_revision": "sha256:5caed90ac831737a11898565002a9789f3f74348cfc2448860ca875bf01a7456", + "limit_seconds": null, + "planned_method_id": "python.check-default-read-order", + "result_candidate_sha256": "ea62d085e2976c0b495ef719e9214110c696bd7fc412f2c7bdc17fe911d5e5d5", + "result_ref": "research/replan-check-defaults/observations.json", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Diagnostic isolates access-order dependence; scoped repair selected without production edits." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:ea62d085e2976c0b495ef719e9214110c696bd7fc412f2c7bdc17fe911d5e5d5", + "comparison_scope": "local:check-shared-default-access-order", + "evaluation_ref": "python:3.11:paired-config-read-edit:v1", + "evidence_ref": "research/replan-check-defaults/observations.json", + "higher_is_better": false, + "input_revision": "sha256:5caed90ac831737a11898565002a9789f3f74348cfc2448860ca875bf01a7456", + "metric": "access_order_divergent_pairs", + "unit": "pairs", + "value": 6 + } + ], + "node_id": "question_ce0f7b93bf30e98248f7bc3b6d4246adad597b92fdd767f6b84c233192aad56a", + "recorded_at": "2026-10-24T00:05:17+00:00", + "result_id": "finding_3ab0d5f8e18f75e3e00621436e57d2e1728d236a0b70ab1b2469980ae43e65b1", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:9b835deba14817e0", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:5caed90ac831737a11898565002a9789f3f74348cfc2448860ca875bf01a7456 Applicability: Candidate 28 on Python 3.11: direct CheckContext object/dictionary construction, with operation/phase resolution and explicit named settings as controls. No native failure attribution. Observation: Six of 15 paired scenarios diverge after a prior read; nine controls agree. Stale statuses reject a 503 response accepted without the read. A separate load_checks confirmation retains seven checks after disabling all; without the first call it selects none. Source is unchanged. Interpretation: The scoped access-order invariant fails. ChecksConfig.get materializes inherited defaults into named cache entries, and later direct context lookups reuse stale values. This affects enabled selection, expected statuses and maximum response time. Operation/phase resolution already treats cached reads differently from overrides, and explicit named controls retain their values. The original inputs remain unchanged. The separate load_checks probe confirms the enabled symptom but overlaps the main cases and is not added to their count. Reuse existing loading/validation/generation audits, whose 12 source files still match. Select a scoped refresh repair preserving named overrides and mutations, independent mutable values and the working operation path; do not erase all cached settings. Candidate 28 remains a measured tied best and all six requirements stay open. The public native report reveals no assertion identities, so this reproduction neither explains its gaps nor predicts recovery.", + "tags": [ + "writeback-result" + ], + "title": "Do direct check-configuration reads preserve the effects of later shared-default edits independently of access order?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "d933d937a239dfec", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-24T00:15:42+00:00", + "result_id": "question_9a6a8466757245afb2f70fd776cbc811246e6f0b4b3722adeb79359b7691218d", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:2d33ec66030367b4", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Archived candidates 28/29 on identical 45-case Python 3.11 inputs, including direct check contexts and operation/phase resolution. Native acceptance and other runtimes are outside this comparison.", + "title": "Does refreshing shared check defaults preserve explicit named edits and operation resolution independently of earlier reads?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "caf369788a532185", + "event_kind": "finding", + "evidence_refs": [ + "research/check-default-refresh/REPORT.md", + "research/check-default-refresh/observations.json", + "research/check-default-refresh/evaluation-input-manifest.json", + "research/check-default-refresh/candidate-29-manifest.json", + "research/check-default-refresh/candidate-diff.patch", + "research/check-default-refresh/test_defaults.py", + "research/check-default-refresh/baseline-execution.json" + ], + "execution_observation": { + "candidate_sha256": "e6197e5600534947751f805086266f992848649b3f2f241ecc72a348b713cfd4", + "change_reason": null, + "elapsed_seconds": 1.2488894169946434, + "evidence_refs": [ + "research/check-default-refresh/baseline-execution.json", + "research/check-default-refresh/candidate-29-manifest.json", + "research/check-default-refresh/candidate-diff.patch", + "research/check-default-refresh/evaluation-input-manifest.json", + "research/check-default-refresh/observations.json", + "research/check-default-refresh/test_defaults.py" + ], + "executed_method_id": "pytest.archived-check-default-regressions", + "exit_code": 0, + "input_revision": "sha256:df217dd9a933646e9a586e77a7a27f57355fe9c02edf0a3e6b80076d1e06190b", + "limit_seconds": null, + "planned_method_id": "pytest.archived-check-default-regressions", + "result_candidate_sha256": "e6197e5600534947751f805086266f992848649b3f2f241ecc72a348b713cfd4", + "result_ref": "research/check-default-refresh/observations.json", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Repair and focused validation pass; preserve candidate for continued integration and eligible native evaluation." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:e6197e5600534947751f805086266f992848649b3f2f241ecc72a348b713cfd4", + "comparison_scope": "local:check-default-refresh", + "evaluation_ref": "python:3.11:pytest:8.4.2:check-default-regressions:v1", + "evidence_ref": "research/check-default-refresh/observations.json", + "higher_is_better": true, + "input_revision": "sha256:df217dd9a933646e9a586e77a7a27f57355fe9c02edf0a3e6b80076d1e06190b", + "metric": "check_default_regressions_passed", + "unit": "tests", + "value": 45 + } + ], + "node_id": "question_9a6a8466757245afb2f70fd776cbc811246e6f0b4b3722adeb79359b7691218d", + "recorded_at": "2026-10-24T00:15:42+00:00", + "result_id": "finding_0f3443babccddea93c8e5bfa855b0f004233c192395244553c4b9ca1281491e3", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:59d244fe85e14d4a", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:df217dd9a933646e9a586e77a7a27f57355fe9c02edf0a3e6b80076d1e06190b Applicability: Archived candidates 28/29 on identical 45-case Python 3.11 inputs, including direct check contexts and operation/phase resolution. Native acceptance and other runtimes are outside this comparison. Observation: Candidate 28 fails 25 and passes 20 cases; candidate 29 passes all 45 and the 690 focused cases. Shared edits take effect after reads while named settings and list mutations remain intact. All 317 candidate files verify. Interpretation: Explicit input fields and assignments are tracked separately from inherited values. Lookup refreshes inherited fields and detects in-place list edits without replacing named configuration objects. Operation reconciliation uses explicit settings, preventing refreshed cache values from becoming overrides. The comparison reuses the 15 original paired scenarios and adds enabled selection, partial overrides, same-value assignments, copying, mutable isolation and repeated resolution across four phases. Source archives and identical inputs verify. Initial and archived repair runs agree; only distinct final cases count. This local result does not identify or predict native failures. Candidate 28 remains preserved at the native best; 29 is unsubmitted during cooldown. Keep all six requirements and known health-check/full-suite/runtime limitations open. Next inspect remaining configuration integration against retained audits, requiring a new reproduction before changes, and evaluate the preserved successor once eligible.", + "tags": [ + "writeback-result" + ], + "title": "Does refreshing shared check defaults preserve explicit named edits and operation resolution independently of earlier reads?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "e396914808b752d0", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-24T00:22:39+00:00", + "result_id": "question_bd42ff6fdbe17cdb377d200f24429ca629f4d5cb0fc52750abe67a576cb44754", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:2d8ce910e3499332", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Candidate 29 on Python 3.11 with four config loading routes and an owned WSGI app. Phase configs are explicitly selected; automatic CLI discovery and stateful execution are outside this scope.", + "title": "Do loaded and discovered configuration values reach generated calls and enabled checks after project, operation and phase precedence?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "58834052d4d08e23", + "event_kind": "finding", + "evidence_refs": [ + "research/config-runtime-integration/REPORT.md", + "research/config-runtime-integration/observations.json", + "research/config-runtime-integration/test_runtime.py", + "research/config-runtime-integration/evaluation-input-manifest.json", + "research/config-runtime-integration/existing-lineage.json", + "research/config-runtime-integration/source-reuse.json", + "research/config-runtime-integration/runtime-provenance.json" + ], + "execution_observation": { + "candidate_sha256": "e6197e5600534947751f805086266f992848649b3f2f241ecc72a348b713cfd4", + "change_reason": null, + "elapsed_seconds": 6.214482419003616, + "evidence_refs": [ + "research/config-runtime-integration/evaluation-input-manifest.json", + "research/config-runtime-integration/existing-lineage.json", + "research/config-runtime-integration/observations.json", + "research/config-runtime-integration/runtime-provenance.json", + "research/config-runtime-integration/source-reuse.json", + "research/config-runtime-integration/test_runtime.py" + ], + "executed_method_id": "pytest.config-wsgi-runtime", + "exit_code": 0, + "input_revision": "sha256:491013dd8daa35454825ba8c9328b73b550daf926644c231583481bb29d589c6", + "limit_seconds": null, + "planned_method_id": "pytest.config-wsgi-runtime", + "result_candidate_sha256": "e6197e5600534947751f805086266f992848649b3f2f241ecc72a348b713cfd4", + "result_ref": "research/config-runtime-integration/observations.json", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "The configuration composition passes; no production correction is justified by this probe. Continue remaining acceptance work." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:e6197e5600534947751f805086266f992848649b3f2f241ecc72a348b713cfd4", + "comparison_scope": "local:config-runtime-integration", + "evaluation_ref": "python:3.11:pytest:8.4.2:config-wsgi-runtime:v1", + "evidence_ref": "research/config-runtime-integration/observations.json", + "higher_is_better": true, + "input_revision": "sha256:491013dd8daa35454825ba8c9328b73b550daf926644c231583481bb29d589c6", + "metric": "config_runtime_cases_passed", + "unit": "tests", + "value": 32 + } + ], + "node_id": "question_bd42ff6fdbe17cdb377d200f24429ca629f4d5cb0fc52750abe67a576cb44754", + "recorded_at": "2026-10-24T00:22:39+00:00", + "result_id": "finding_64e9f71a798cf9a1149c65fd15e3b732888e668ce8d5f889c228ea9bc41c16af", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:da3da56914bf64d9", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:491013dd8daa35454825ba8c9328b73b550daf926644c231583481bb29d589c6 Applicability: Candidate 29 on Python 3.11 with four config loading routes and an owned WSGI app. Phase configs are explicitly selected; automatic CLI discovery and stateful execution are outside this scope. Observation: All 32 runtime cases and nine supported existing loader cases pass on unchanged candidate 29. Precedence, generated modes/counts, deterministic calls, headers, unified responses and check selection agree. The initial removed-wrapper collection error is preserved. Interpretation: No defect is reproduced in this composition. The matrix combines dictionary/string/file/discovery loading with matching and nonmatching operation overrides across four phase configurations. Generated tests call an owned WSGI app and verify expected statuses, rejection controls and mutable isolation; repeated examples are not extra test counts. Nine legacy assertions retain their meaning after namespace migration. An initial import of removed runner.prepare prevented collection; seven parameters for that obsolete wrapper are excluded and remain unexecuted, with their exact failed input retained. Source and input hashes verify. Ten configuration files match the old audits; two changed with candidate 29, so historical findings stay separately attributed. No native or full-suite success is inferred. Keep the current candidate and all six requirements open; next inspect remaining Response/transport boundaries against existing evidence and submit only when eligible.", + "tags": [ + "writeback-result" + ], + "title": "Do loaded and discovered configuration values reach generated calls and enabled checks after project, operation and phase precedence?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "7e1e4da053179a04", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-24T00:36:27+00:00", + "result_id": "question_a67989ebac8265861951bfb4b1d2991c0a9cdba98be84db8c131eed9764b12f5", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:61d45dcbfd8579e3", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Python 3.11 with Werkzeug 3.1.8, owned HTTP/WSGI/ASGI endpoints and the retained redirect matrix. Native-client behavior is preserved; arbitrary custom redirect implementations are outside scope.", + "title": "Do WSGI client options and redirected prepared requests preserve the actual exchange through normalized responses and reports?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "1805a254682b506d", + "event_kind": "finding", + "evidence_refs": [ + "research/redirect-request-recording/REPORT.md", + "research/redirect-request-recording/observations.json", + "research/redirect-request-recording/evaluation-input-manifest.json", + "research/redirect-request-recording/candidate-diff.patch", + "research/redirect-request-recording/candidate29-execution.json", + "research/redirect-request-recording/candidate30-execution.json", + "research/redirect-request-recording/lineage.json" + ], + "execution_observation": { + "candidate_sha256": "d79865a8d18edaad3b47bb483e0bf904d53241895d07fcc539d8777fe19aeb44", + "change_reason": null, + "elapsed_seconds": 2.108995959002641, + "evidence_refs": [ + "research/redirect-request-recording/candidate-diff.patch", + "research/redirect-request-recording/candidate29-execution.json", + "research/redirect-request-recording/candidate30-execution.json", + "research/redirect-request-recording/evaluation-input-manifest.json", + "research/redirect-request-recording/lineage.json", + "research/redirect-request-recording/observations.json" + ], + "executed_method_id": "pytest.redirect-client-controls", + "exit_code": 0, + "input_revision": "sha256:784da08e0329d0169b4b49d5dcea9fdf93334a288592ed0719864bac2226b403", + "limit_seconds": null, + "planned_method_id": "pytest.redirect-client-controls", + "result_candidate_sha256": "d79865a8d18edaad3b47bb483e0bf904d53241895d07fcc539d8777fe19aeb44", + "result_ref": "research/redirect-request-recording/observations.json", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "The controlled redirect scope passes. All-six acceptance and native evaluation remain open." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:d79865a8d18edaad3b47bb483e0bf904d53241895d07fcc539d8777fe19aeb44", + "comparison_scope": "local:redirect-request-recording", + "evaluation_ref": "python:3.11:pytest:8.4.2:werkzeug-redirect-controls:v1", + "evidence_ref": "research/redirect-request-recording/observations.json", + "higher_is_better": true, + "input_revision": "sha256:784da08e0329d0169b4b49d5dcea9fdf93334a288592ed0719864bac2226b403", + "metric": "redirect_cases_passed", + "unit": "tests", + "value": 77 + } + ], + "node_id": "question_a67989ebac8265861951bfb4b1d2991c0a9cdba98be84db8c131eed9764b12f5", + "recorded_at": "2026-10-24T00:36:27+00:00", + "result_id": "finding_ceaa84ab9dbce0ef03760167007869ebd493c4b3939431767674422a944f0adf", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:6d1abc525a0d033d", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:784da08e0329d0169b4b49d5dcea9fdf93334a288592ed0719864bac2226b403 Applicability: Python 3.11 with Werkzeug 3.1.8, owned HTTP/WSGI/ASGI endpoints and the retained redirect matrix. Native-client behavior is preserved; arbitrary custom redirect implementations are outside scope. Observation: Candidate 29 fails 44 of 77 cases; candidate 30 passes all 77. WSGI client options reach Client.open and recorded bodies match final endpoints. Another 22 CLI and 690 focused cases pass. Initial assumptions and intermediate failures are preserved. Interpretation: Two WSGI integration defects are resolved: client-only options reached EnvironBuilder, and redirect responses retained the original upload in their prepared request. The fix forwards buffered/follow_redirects separately and uses the final request length to record the remaining encoded body. Native Werkzeug controls show that consumed input need not replay on 307/308; six intermediate failures reflected an overly strong test assumption, not a framework delivery defect. Exact earlier inputs, the partial repair and failures remain retained. Final controlled runs use identical inputs and archived source with in-process origin checks. Report serialization, method changes, mixed chains, omitted bodies, partial consumption and caller clients pass. The core Response audit is reused as historical evidence, not rerun. No native evaluation or score improvement is inferred. Keep all six requirements open; next continue integration work and evaluate the preserved successor only when eligible.", + "tags": [ + "writeback-result" + ], + "title": "Do WSGI client options and redirected prepared requests preserve the actual exchange through normalized responses and reports?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "09c05a4e1fd61d6e", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-24T00:47:50+00:00", + "result_id": "question_5e9d6886e711816d18b798ac69d3b9286920219406f806a89859aaaf08d7ba0b", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:cba7d61c964becf1", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Python 3.11; constructor and requests/HTTPX/Werkzeug responses, owned HTTP/WSGI/ASGI endpoints, public after_call hooks, OpenAPI CLI checks and GraphQL validation.", + "title": "Do response checks and reports observe after_call body replacement when an earlier hook has populated JSON and base64 caches?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "38bb4941dc1a1493", + "event_kind": "finding", + "evidence_refs": [ + "research/response-hook-cache/REPORT.md", + "research/response-hook-cache/observations.json", + "research/response-hook-cache/evaluation-input-manifest.json", + "research/response-hook-cache/candidate-diff.patch", + "research/response-hook-cache/candidate30-execution.json", + "research/response-hook-cache/candidate31-execution.json", + "research/response-hook-cache/lineage.json" + ], + "execution_observation": { + "candidate_sha256": "8c593f98f97ed7b7cddd9f66efd189e73556d3b9bc32211c5cc90f233181faa9", + "change_reason": null, + "elapsed_seconds": 2.2993262090021744, + "evidence_refs": [ + "research/response-hook-cache/candidate-diff.patch", + "research/response-hook-cache/candidate30-execution.json", + "research/response-hook-cache/candidate31-execution.json", + "research/response-hook-cache/evaluation-input-manifest.json", + "research/response-hook-cache/lineage.json", + "research/response-hook-cache/observations.json" + ], + "executed_method_id": "pytest.response-hook-cache", + "exit_code": 0, + "input_revision": "sha256:efe13d57006adaac67a8cbce1f5b53d7a0a6e84c95d09b7e7cda8a32c5b9f8f7", + "limit_seconds": null, + "planned_method_id": "pytest.response-hook-cache", + "result_candidate_sha256": "8c593f98f97ed7b7cddd9f66efd189e73556d3b9bc32211c5cc90f233181faa9", + "result_ref": "research/response-hook-cache/observations.json", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Controlled hook/cache checks pass; all-six acceptance and native evaluation remain open." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:8c593f98f97ed7b7cddd9f66efd189e73556d3b9bc32211c5cc90f233181faa9", + "comparison_scope": "local:response-hook-cache", + "evaluation_ref": "python:3.11:pytest:8.4.2:hook-cache-matrix:v1", + "evidence_ref": "research/response-hook-cache/observations.json", + "higher_is_better": true, + "input_revision": "sha256:efe13d57006adaac67a8cbce1f5b53d7a0a6e84c95d09b7e7cda8a32c5b9f8f7", + "metric": "hook_cache_cases_passed", + "unit": "tests", + "value": 82 + } + ], + "node_id": "question_5e9d6886e711816d18b798ac69d3b9286920219406f806a89859aaaf08d7ba0b", + "recorded_at": "2026-10-24T00:47:50+00:00", + "result_id": "finding_6359b7936c60b71b0dbf3e6a932f72132bd359cc9f19a367d74031b2563ab3b7", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:8b767f5886ef38ae", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:efe13d57006adaac67a8cbce1f5b53d7a0a6e84c95d09b7e7cda8a32c5b9f8f7 Applicability: Python 3.11; constructor and requests/HTTPX/Werkzeug responses, owned HTTP/WSGI/ASGI endpoints, public after_call hooks, OpenAPI CLI checks and GraphQL validation. Observation: Candidate 30 fails 57 of 82 hook/cache cases; candidate 31 passes all 82 plus 773 regressions. Body replacement now invalidates both caches, so OpenAPI/GraphQL checks reject invalid replacements and cassettes retain updated bytes. Interpretation: A documented in-place hook mutation exposed a gap not covered by historical factory audits. Earlier JSON reads concealed later invalid response bodies from both OpenAPI and GraphQL checks; earlier base64 reads left cassette evidence stale. The repair invalidates both caches whenever content is assigned, preserving slots, encoding and normal cache reuse. Identical controlled inputs run against archived candidates with in-process origin checks. Six existing hook cases preserve original assertions with isolated fixtures; 77 redirect cases and 690 focused cases also pass. Native history remains seven completed submissions, last tied best 0.786257, with cooldown still active at the eligibility read. No native improvement or full task completion is inferred. The scoped hook/cache defect is resolved; all six targets, known health-check/full-suite/minimum-runtime gaps and native validation remain open. Next reconcile eligibility and evaluate the preserved candidate, retaining source/result identity and native-best alternatives.", + "tags": [ + "writeback-result" + ], + "title": "Do response checks and reports observe after_call body replacement when an earlier hook has populated JSON and base64 caches?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "4b9407441ceece71", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-24T01:00:26+00:00", + "result_id": "question_1f86fb101993293e8a7023711b4748b265cc3798090b99fa6e876437db135dfc", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:e4e6365e869ccd13", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Public native report c5941a995ee5 bound to candidate 31; phase totals only. Earlier native/local results retain separate source identities and unverified evaluator comparability.", + "title": "Does preserved candidate 31 satisfy all six roadmap phases in native submission c5941a995ee5?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "691c8d179bf5a474", + "event_kind": "finding", + "evidence_refs": [ + "research/native-candidate-31/REPORT.md", + "research/native-candidate-31/native-result-second.txt", + "research/native-candidate-31/submission-source.json", + "research/native-candidate-31/native-summary.json", + "research/native-candidate-31/source-comparison.json", + "research/native-candidate-31/candidate28-to-31.patch", + "research/native-candidate-31/local-evidence-lineage.json", + "research/native-candidate-31/next-route.json" + ], + "execution_observation": { + "candidate_sha256": "8c593f98f97ed7b7cddd9f66efd189e73556d3b9bc32211c5cc90f233181faa9", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "research/native-candidate-31/native-result-second.txt", + "research/native-candidate-31/submission-source.json" + ], + "executed_method_id": "sforge-submit.native", + "exit_code": 0, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "limit_seconds": null, + "planned_method_id": "sforge-submit.native", + "result_candidate_sha256": "8c593f98f97ed7b7cddd9f66efd189e73556d3b9bc32211c5cc90f233181faa9", + "result_ref": "research/native-candidate-31/native-result-second.txt", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "The accepted result ties best; five phases remain incomplete. Continue with the evidence-linked CLI regression route." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:8c593f98f97ed7b7cddd9f66efd189e73556d3b9bc32211c5cc90f233181faa9", + "comparison_scope": "native:c5941a995ee5:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-31/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "weighted_score", + "unit": "ratio", + "value": 0.786257 + }, + { + "candidate_ref": "sha256:8c593f98f97ed7b7cddd9f66efd189e73556d3b9bc32211c5cc90f233181faa9", + "comparison_scope": "native:c5941a995ee5:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-31/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "complete_phases", + "unit": "phases", + "value": 1 + }, + { + "candidate_ref": "sha256:8c593f98f97ed7b7cddd9f66efd189e73556d3b9bc32211c5cc90f233181faa9", + "comparison_scope": "native:c5941a995ee5:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-31/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "configuration_checks_passed", + "unit": "checks", + "value": 19 + }, + { + "candidate_ref": "sha256:8c593f98f97ed7b7cddd9f66efd189e73556d3b9bc32211c5cc90f233181faa9", + "comparison_scope": "native:c5941a995ee5:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-31/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "api_checks_passed", + "unit": "checks", + "value": 26 + }, + { + "candidate_ref": "sha256:8c593f98f97ed7b7cddd9f66efd189e73556d3b9bc32211c5cc90f233181faa9", + "comparison_scope": "native:c5941a995ee5:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-31/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "metrics_checks_passed", + "unit": "checks", + "value": 14 + }, + { + "candidate_ref": "sha256:8c593f98f97ed7b7cddd9f66efd189e73556d3b9bc32211c5cc90f233181faa9", + "comparison_scope": "native:c5941a995ee5:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-31/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "transport_checks_passed", + "unit": "checks", + "value": 14 + }, + { + "candidate_ref": "sha256:8c593f98f97ed7b7cddd9f66efd189e73556d3b9bc32211c5cc90f233181faa9", + "comparison_scope": "native:c5941a995ee5:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-31/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "response_checks_passed", + "unit": "checks", + "value": 15 + }, + { + "candidate_ref": "sha256:8c593f98f97ed7b7cddd9f66efd189e73556d3b9bc32211c5cc90f233181faa9", + "comparison_scope": "native:c5941a995ee5:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/native-candidate-31/native-result-second.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "checks_checks_passed", + "unit": "checks", + "value": 16 + } + ], + "node_id": "question_1f86fb101993293e8a7023711b4748b265cc3798090b99fa6e876437db135dfc", + "recorded_at": "2026-10-24T01:00:26+00:00", + "result_id": "finding_ac1d100d6a80aa262b1c8390b1c2a808262a39be1c814949ef2e1912f20686d0", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:97031855a94b36e2", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89 Applicability: Public native report c5941a995ee5 bound to candidate 31; phase totals only. Earlier native/local results retain separate source identities and unverified evaluator comparability. Observation: Candidate 31 scores 0.786257 with 1/6 phases complete: configuration 19/30, API 26/26, metrics 14/15, transport 14/19, Response 15/18, checks 16/18. All totals match candidate 28. Retain both and earlier bests. Interpretation: Five phases remain incomplete, so the scoped acceptance question has a negative answer. One accepted ID binds the completed result to 346 verified candidate files and 179 Python source files; pending feedback is retained. Earlier native-best archives remain recoverable. Undisclosed evaluator/data revisions prevent a controlled causal delta. Local check-default, WSGI and cache fixes are independently supported but have not increased reported native totals. The self-repair review found no control-plane fault; it corrected investigation selection. Existing configuration audits already cover proposed probes. Root test collection stops at missing aiohttp; a 12-case isolated module passes but repeats prior candidate-22 coverage, so it is not new acceptance evidence. Next run the complete original CLI metric-extension module with isolated OpenAPI/GraphQL fixtures, preserving assertions and cleanup, then investigate real failures. Retain all six requirements and known health-check/full-suite/minimum-runtime limits.", + "tags": [ + "writeback-result" + ], + "title": "Does preserved candidate 31 satisfy all six roadmap phases in native submission c5941a995ee5?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "505f3923be077276", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-24T01:11:16+00:00", + "result_id": "question_b7527072dd43881a9a3e14bbac79eb30cc1132d5d9ce7188f263b73218816e0c", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:b3c1c0df385e1f71", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Candidate 31 on Python 3.11; original CLI/ctx helpers, owned OpenAPI HTTP/WSGI and Flask GraphQL endpoints. Includes worker variants and failure recovery; excludes full-suite/native acceptance.", + "title": "Do the original CLI metric extension tests preserve custom OpenAPI/GraphQL callbacks and cleanup through the v4 metric API?" + }, + { + "agent_id": "schema-astra-loopx-20261023-18h02-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "dcc8d3f207c619d6", + "event_kind": "finding", + "evidence_refs": [ + "research/metric-cli-integration/REPORT.md", + "research/metric-cli-integration/observations.json", + "research/metric-cli-integration/input-manifest.json", + "research/metric-cli-integration/migration.json", + "research/metric-cli-integration/migrated.txt", + "research/metric-cli-integration/negative-final.txt", + "research/metric-cli-integration/migrated-execution.json", + "research/metric-cli-integration/negative-final-execution.json" + ], + "execution_observation": { + "candidate_sha256": "8c593f98f97ed7b7cddd9f66efd189e73556d3b9bc32211c5cc90f233181faa9", + "change_reason": null, + "elapsed_seconds": 3.5449651679955423, + "evidence_refs": [ + "research/metric-cli-integration/input-manifest.json", + "research/metric-cli-integration/migrated-execution.json", + "research/metric-cli-integration/negative-final-execution.json", + "research/metric-cli-integration/observations.json" + ], + "executed_method_id": "pytest.cli-metric-integration", + "exit_code": 0, + "input_revision": "sha256:5e865f9756a17f2f8bd2aefce59217fc587bc7e6b5805407d9a23a652f110b9b", + "limit_seconds": null, + "planned_method_id": "pytest.cli-metric-integration", + "result_candidate_sha256": "8c593f98f97ed7b7cddd9f66efd189e73556d3b9bc32211c5cc90f233181faa9", + "result_ref": "research/metric-cli-integration/observations.json", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Scoped integration passes; retain candidate 31 and continue remaining check/configuration acceptance work." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:8c593f98f97ed7b7cddd9f66efd189e73556d3b9bc32211c5cc90f233181faa9", + "comparison_scope": "local:cli-metric-extension-and-recovery", + "evaluation_ref": "python:3.11:pytest:8.4.2:isolated-cli-metrics:v1", + "evidence_ref": "research/metric-cli-integration/observations.json", + "higher_is_better": true, + "input_revision": "sha256:5e865f9756a17f2f8bd2aefce59217fc587bc7e6b5805407d9a23a652f110b9b", + "metric": "integration_cases_passed", + "unit": "tests", + "value": 15 + } + ], + "node_id": "question_b7527072dd43881a9a3e14bbac79eb30cc1132d5d9ce7188f263b73218816e0c", + "recorded_at": "2026-10-24T01:11:16+00:00", + "result_id": "finding_0f908c92bb4651d3475af5013312516c33d23a3022cc5109e962d02467e12281", + "run_id": "checkpoint:eac760c7b2651654d29dee628d8918d23177643d4f5706fc5bd81845299a9a0a:ca2a1afdfb69c298", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:5e865f9756a17f2f8bd2aefce59217fc587bc7e6b5805407d9a23a652f110b9b Applicability: Candidate 31 on Python 3.11; original CLI/ctx helpers, owned OpenAPI HTTP/WSGI and Flask GraphQL endpoints. Includes worker variants and failure recovery; excludes full-suite/native acceptance. Observation: All 15 distinct cases pass: 3 complete original extension tests, 8 existing response-time worker/transport cases and 4 negative/recovery cases. Registry and CLI choices reset, servers stop, and all 346 candidate files remain unchanged. Interpretation: The scoped CLI migration question is answered positively without a production edit. Original tests and assertions are retained; only the removed target decorator changes to metric. The original return-value fixture is weak, so it is not treated as new decorator-identity proof. Original CLI/ctx helpers and GraphQL app/schema are preserved; owned OpenAPI success endpoints replace optional root fixtures. Initial setup errors came from a pytest plugin-name collision; two later failures came from a new stdout assertion for stderr usage errors. Both initial inputs/outputs remain retained. Corrected negative tests verify rejection and successful recovery after callback exceptions. No new native evaluation occurred, and local passes do not close the five incomplete native phases. Keep candidate 31 and earlier bests; next run original CLI check-selection/configured-status regressions with snapshot expectations preserved. Full-suite, health-check and minimum-runtime limits remain open.", + "tags": [ + "writeback-result" + ], + "title": "Do the original CLI metric extension tests preserve custom OpenAPI/GraphQL callbacks and cleanup through the v4 metric API?" + }, + { + "agent_id": "schema-7269-test-solmax-r3-20261025-4h05-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "254d11022e9b15cd", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-25T15:07:27+00:00", + "result_id": "question_87c16c8085b2262004547ad1633a1dfb137f40f8b729c04802c2ad15e80ea0ce", + "run_id": "explore-result:74d128f7df03fc53d3716b03ef0fc96cecddc81aa89702f14ae837f6e3d39a4d", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Candidate 31, Python 3.11, isolated original CLI helpers, owned HTTP/ASGI endpoints; 18 original cases plus 8 lifecycle cases. Local evidence only; no native or full-suite completion claim.", + "title": "How do the complete historical CLI-check oracles and custom-check lifecycle behave on archived candidate 31?" + }, + { + "agent_id": "schema-7269-test-solmax-r3-20261025-4h05-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "bca33c5c593b6f2c", + "event_kind": "finding", + "evidence_refs": [ + "research/cli-check-integration/REPORT.md", + "research/cli-check-integration/combined-original-and-selection-case-accounting.json", + "research/cli-check-integration/input-manifest.json", + "research/cli-check-integration/evaluated-input-manifest.json", + "research/cli-check-integration/migration.json", + "research/cli-check-integration/combined-original-and-selection-execution.json", + "research/cli-check-integration/combined-original-and-selection-snapshot-diffs.json", + "research/cli-check-integration/environment-evidence.json" + ], + "execution_observation": { + "candidate_sha256": "8c593f98f97ed7b7cddd9f66efd189e73556d3b9bc32211c5cc90f233181faa9", + "change_reason": null, + "elapsed_seconds": 5.5642102519923355, + "evidence_refs": [ + "research/cli-check-integration/combined-original-and-selection-case-accounting.json", + "research/cli-check-integration/combined-original-and-selection-execution.json", + "research/cli-check-integration/evaluated-input-manifest.json" + ], + "executed_method_id": "pytest.cli-check-original-and-selection", + "exit_code": 1, + "input_revision": "sha256:442db449df6630d3daaaef3da033ee4c7d00b6a6510244bcee2ce198bca01d91", + "limit_seconds": null, + "planned_method_id": "pytest.cli-check-original-and-selection", + "result_candidate_sha256": "8c593f98f97ed7b7cddd9f66efd189e73556d3b9bc32211c5cc90f233181faa9", + "result_ref": "research/cli-check-integration/combined-original-and-selection-case-accounting.json", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "All original cases executed and four output mismatches accounted for. Keep Todo open for semantic route review; production unchanged." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:8c593f98f97ed7b7cddd9f66efd189e73556d3b9bc32211c5cc90f233181faa9", + "comparison_scope": "local:cli-checks-original-and-selection:v1", + "evaluation_ref": "pytest:8.4.2:python:3.11:isolated-cli-checks:v1", + "evidence_ref": "research/cli-check-integration/combined-original-and-selection-case-accounting.json", + "higher_is_better": true, + "input_revision": "sha256:442db449df6630d3daaaef3da033ee4c7d00b6a6510244bcee2ce198bca01d91", + "metric": "cli_check_cases_passed", + "unit": "tests", + "value": 22 + }, + { + "candidate_ref": "sha256:8c593f98f97ed7b7cddd9f66efd189e73556d3b9bc32211c5cc90f233181faa9", + "comparison_scope": "local:cli-checks-original-and-selection:v1", + "evaluation_ref": "pytest:8.4.2:python:3.11:isolated-cli-checks:v1", + "evidence_ref": "research/cli-check-integration/combined-original-and-selection-case-accounting.json", + "higher_is_better": false, + "input_revision": "sha256:442db449df6630d3daaaef3da033ee4c7d00b6a6510244bcee2ce198bca01d91", + "metric": "cli_check_cases_failed", + "unit": "tests", + "value": 4 + } + ], + "node_id": "question_87c16c8085b2262004547ad1633a1dfb137f40f8b729c04802c2ad15e80ea0ce", + "recorded_at": "2026-10-25T15:07:27+00:00", + "result_id": "finding_74d128f7df03fc53d3716b03ef0fc96cecddc81aa89702f14ae837f6e3d39a4d", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:442db449df6630d3daaaef3da033ee4c7d00b6a6510244bcee2ce198bca01d91 Applicability: Candidate 31, Python 3.11, isolated original CLI helpers, owned HTTP/ASGI endpoints; 18 original cases plus 8 lifecycle cases. Local evidence only; no native or full-suite completion claim. Observation: All 26 cases execute: 22 pass and 4 original snapshots fail. Ten of 14 goldens match unchanged. Exact output differences are accounted for; registry/choices reset and owned servers stop. Candidate 31 is unchanged. Interpretation: The complete CLI frontier is now executed, including all seven original functions, 18 cases and 14 snapshots with unchanged assertions. Four literal failures remain: embedded historical port and failure-block order; added configured-status detail; two Click 8.4.1 stderr/help-hint changes. Explicit diagnostic comparisons account for every other byte but never override pytest outcomes. Eight custom-check selection, failure/error recovery, unknown-name, replacement and unregister/reset cases pass separately and in the combined process. Import origins and all 346 candidate files verify. No production defect is demonstrated, so no successor or unchanged native submission is justified. Keep the Todo and all six requirements open; review this route rather than blindly rerun. Known health-check/full-suite/minimum-runtime limits remain.", + "tags": [ + "writeback-result" + ], + "title": "How do the complete historical CLI-check oracles and custom-check lifecycle behave on archived candidate 31?" + }, + { + "agent_id": "schema-7269-test-solmax-r3-20261025-4h05-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "b913150d7045d5ee", + "event_kind": "node", + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "node_kind": "question", + "recorded_at": "2026-10-25T15:39:04+00:00", + "result_id": "question_d539ed62b54514d9626df6c4a4fab26d2cd84e47da329c2054cc45de2122957f", + "run_id": "explore-result:7f6b4484ff86dac1b15bab6c35bfdf660c24dd0638a81cbee4e1eacd25b1963d", + "schema_version": "loopx_explore_result_event_v0", + "status": "resolved", + "summary": "Public native b16cda9c8fac bound to archived 32; local paired lifecycle and integration evidence is separate. Aggregate native phases only; no hidden assertions or controlled cross-round attribution.", + "title": "Does candidate 32's live CLI registry repair satisfy all six roadmap phases in native submission b16cda9c8fac?" + }, + { + "agent_id": "schema-7269-test-solmax-r3-20261025-4h05-agent", + "boundary": { + "absolute_paths_recorded": false, + "credential_values_recorded": false, + "raw_logs_recorded": false, + "raw_session_transcript_recorded": false, + "raw_task_text_recorded": false, + "raw_trajectory_recorded": false + }, + "event_id": "5c1905290a195831", + "event_kind": "finding", + "evidence_refs": [ + "research/cli-registry-lifecycle/REPORT.md", + "research/cli-registry-lifecycle/observations.json", + "research/cli-registry-lifecycle/input-manifest.json", + "research/cli-registry-lifecycle/candidate31-to-32.patch", + "research/cli-registry-lifecycle/native-result-first.txt", + "research/cli-registry-lifecycle/submission-source.json", + "research/cli-registry-lifecycle/native-summary.json", + "research/cli-registry-lifecycle/integrated32-case-accounting.json" + ], + "execution_observation": { + "candidate_sha256": "b047c2641578669da9469f365fef32edba65fe2328fcb849b78bc4c03dde7f53", + "change_reason": null, + "elapsed_seconds": null, + "evidence_refs": [ + "research/cli-registry-lifecycle/native-result-first.txt", + "research/cli-registry-lifecycle/submission-source.json" + ], + "executed_method_id": "sforge-submit.native", + "exit_code": 0, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "limit_seconds": null, + "planned_method_id": "sforge-submit.native", + "result_candidate_sha256": "b047c2641578669da9469f365fef32edba65fe2328fcb849b78bc4c03dde7f53", + "result_ref": "research/cli-registry-lifecycle/native-result-first.txt", + "schema_version": "explore_execution_observation_v0", + "search_iterations": null, + "stop_reason": "Final native result ties reported best with five phases open. Retain the locally validated repair and continue through the claimed remaining-contract successor." + }, + "goal_id": "schema-7269-test-solmax-r3-20261025-4h05", + "measurements": [ + { + "candidate_ref": "sha256:b047c2641578669da9469f365fef32edba65fe2328fcb849b78bc4c03dde7f53", + "comparison_scope": "native:b16cda9c8fac:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/cli-registry-lifecycle/native-result-first.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "weighted_score", + "unit": "ratio", + "value": 0.786257 + }, + { + "candidate_ref": "sha256:b047c2641578669da9469f365fef32edba65fe2328fcb849b78bc4c03dde7f53", + "comparison_scope": "native:b16cda9c8fac:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/cli-registry-lifecycle/native-result-first.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "complete_phases", + "unit": "phases", + "value": 1 + }, + { + "candidate_ref": "sha256:b047c2641578669da9469f365fef32edba65fe2328fcb849b78bc4c03dde7f53", + "comparison_scope": "native:b16cda9c8fac:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/cli-registry-lifecycle/native-result-first.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "configuration_checks_passed", + "unit": "checks", + "value": 19 + }, + { + "candidate_ref": "sha256:b047c2641578669da9469f365fef32edba65fe2328fcb849b78bc4c03dde7f53", + "comparison_scope": "native:b16cda9c8fac:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/cli-registry-lifecycle/native-result-first.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "api_checks_passed", + "unit": "checks", + "value": 26 + }, + { + "candidate_ref": "sha256:b047c2641578669da9469f365fef32edba65fe2328fcb849b78bc4c03dde7f53", + "comparison_scope": "native:b16cda9c8fac:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/cli-registry-lifecycle/native-result-first.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "metrics_checks_passed", + "unit": "checks", + "value": 14 + }, + { + "candidate_ref": "sha256:b047c2641578669da9469f365fef32edba65fe2328fcb849b78bc4c03dde7f53", + "comparison_scope": "native:b16cda9c8fac:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/cli-registry-lifecycle/native-result-first.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "transport_checks_passed", + "unit": "checks", + "value": 14 + }, + { + "candidate_ref": "sha256:b047c2641578669da9469f365fef32edba65fe2328fcb849b78bc4c03dde7f53", + "comparison_scope": "native:b16cda9c8fac:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/cli-registry-lifecycle/native-result-first.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "response_checks_passed", + "unit": "checks", + "value": 15 + }, + { + "candidate_ref": "sha256:b047c2641578669da9469f365fef32edba65fe2328fcb849b78bc4c03dde7f53", + "comparison_scope": "native:b16cda9c8fac:unversioned-evaluator", + "evaluation_ref": "sforge-submit:roadmap_structured_json", + "evidence_ref": "research/cli-registry-lifecycle/native-result-first.txt", + "higher_is_better": true, + "input_revision": "sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89", + "metric": "checks_checks_passed", + "unit": "checks", + "value": 16 + } + ], + "node_id": "question_d539ed62b54514d9626df6c4a4fab26d2cd84e47da329c2054cc45de2122957f", + "recorded_at": "2026-10-25T15:39:04+00:00", + "result_id": "finding_7f6b4484ff86dac1b15bab6c35bfdf660c24dd0638a81cbee4e1eacd25b1963d", + "schema_version": "loopx_explore_result_event_v0", + "status": "confirmed", + "summary": "Input revision: sha256:f9d3cc5eb46674c6280da6e43a746c1ce3b18968cc39124d9a5b3ae979ebad89 Applicability: Public native b16cda9c8fac bound to archived 32; local paired lifecycle and integration evidence is separate. Aggregate native phases only; no hidden assertions or controlled cross-round attribution. Observation: Candidate 32 scores 0.786257, 1/6 phases: configuration 19/30, API 26/26, metrics 14/15, transport 14/19, Response 15/18, checks 16/18. Eleven local matched cases fail on 31/pass on 32; 692 regressions pass and four historical snapshots remain failing. Interpretation: The scoped native acceptance question has a negative answer. The current accepted ID directly binds its final feedback to 352 verified candidate files and all 179 source Python files. Reported native totals match earlier best summaries; undisclosed evaluator/data/resource versions prevent a causal native delta. Separate fixed-input local comparisons prove a real registry lifecycle repair: removed names are rejected, direct registrations are selectable, and help/completion read current registries. Candidate 32 has 729 distinct local cases: 725 pass, four unchanged historical output failures remain. Retain 32 alongside 31 and earlier bests. The finite CLI task is superseded with a claimed implementation successor preserving all six targets, remaining native gaps, original goldens and health-check/full-suite/minimum-runtime limits. Begin a distinct uncovered configuration contract using existing audits, not repeated tests or another unchanged submission.", + "tags": [ + "writeback-result" + ], + "title": "Does candidate 32's live CLI registry repair satisfy all six roadmap phases in native submission b16cda9c8fac?" + } + ] +} diff --git a/examples/personal-workspace-browser/research-history.md b/examples/personal-workspace-browser/research-history.md index 407682ca31..9581fcde63 100644 --- a/examples/personal-workspace-browser/research-history.md +++ b/examples/personal-workspace-browser/research-history.md @@ -2,8 +2,8 @@ The default read-only preview uses the complete saved Explore log from the 3 November Grid Turing Robot context-on run. The `research-real-benchmark` -browser scenario reads both context arms independently through the real HTTP -history owner. It verifies complete original text and append order, download +browser scenario reads both context arms, the Ad and Schema histories below, +and the saved Graph analysis through the real HTTP history owner. It verifies complete original text and append order, download readback, source preservation, no writes, and desktop/390px reading. These are the core populated-page references; synthetic fixtures cover boundaries such as races, artificial paging pressure and missing fields. @@ -46,6 +46,34 @@ original JSONL digests and sizes. The two source byte totals are 80,992 and 20,575; their exact original serialization can be reconstructed as described below. No model or solver is called by this scenario. +The same core scenario now includes two record-rich histories from different +tasks, with all **349 original events** retained: + +| Source | Records | Node / finding records | Recording interval (UTC) | Original JSONL SHA-256 | +| --- | ---: | ---: | --- | --- | +| [Ad placement](../../apps/presentation/dashboard/src/data/fixtures/edgebench-ad-20261023-history.json) | 255 | 65 / 190 | 22 October 17:31:31–23 October 07:35:25 | `23a55562f8ad9b808a5633c479451b78b13ad5370a35190d04fab407e3824a72` | +| [Schemathesis](../../apps/presentation/dashboard/src/data/fixtures/edgebench-schema-20261025-history.json) | 94 | 47 / 47 | 23 October 17:50:32–25 October 15:39:04 | `5473191a1f23537a48f38ba82d0dd8ba8ed5ad7daffd944f28910e7d6e19e0ab` | + +Selection combines recency with the reading problem. The inspected 26 October +Ad and Schema captures contain only two and four events. These richer captures +exercise real pagination, long evidence and updates: one Ad question has **39 +recorded versions**; its 65 node events describe three distinct questions. A +confirmed finding rejects a variant after a negative local comparison. The +reader must preserve that original conclusion and make every version reachable. +The Schema capture retains inherited records and originating Agent/run identities +across three UTC dates; that recording span does not measure the destination +solver's active time or invent a parent/child relationship. + +Neither history records explicit edges or parent links. Ad remains +`runner_invalid`; Schema remains `cancelled`; both are score-uncountable. No +events were sampled, rewritten or redacted. The original byte sizes are 396,701 +and 271,025. The core journey reads all display pages, exports every original +record, compares the first and last of the 39 Ad versions and returns keyboard +focus to the selected record at desktop/390px widths. Pass either fixture to +the same preview's `--fixture` option; their original Goal IDs stay intact. +These display captures do not include their complete session trajectories or +qualify installed App adoption, paid analysis or a scientific comparison. + ## Earlier research checkpoints ### Saved analysis reference diff --git a/examples/personal-workspace-browser/research-real-benchmark.mjs b/examples/personal-workspace-browser/research-real-benchmark.mjs index 6d085a0560..0253a115d0 100644 --- a/examples/personal-workspace-browser/research-real-benchmark.mjs +++ b/examples/personal-workspace-browser/research-real-benchmark.mjs @@ -73,15 +73,20 @@ async function readSavedGraphAnalysis(browser, collectCoverage) { } } -// Core whole-page reference: unmodified display state from two real terminal -// runs. Synthetic scenarios remain separate boundary/race/overflow tests. +// Core whole-page references: complete original display state from three tasks. +// Synthetic scenarios remain separate boundary/race/overflow tests. export const researchRealBenchmarkScenario = { id: "research-real-benchmark", async run({browser, collectCoverage}) { const coverageEntries = []; const observations = []; - for (const arm of ["on", "off"]) { - const path = resolve(repoRoot, `apps/presentation/dashboard/src/data/fixtures/edgebench-grid-context-${arm}-20261103-history.json`); + for (const [arm, fixture] of [ + ["grid-on", "edgebench-grid-context-on-20261103-history.json"], + ["grid-off", "edgebench-grid-context-off-20261103-history.json"], + ["ad", "edgebench-ad-20261023-history.json"], + ["schema", "edgebench-schema-20261025-history.json"], + ]) { + const path = resolve(repoRoot, "apps/presentation/dashboard/src/data/fixtures", fixture); const capture = JSON.parse(await readFile(path, "utf8")); const owner = await startHistoryOwner(path); let page; @@ -89,7 +94,7 @@ export const researchRealBenchmarkScenario = { page = await browser.newPage({viewport: {width: 1440, height: 900}}); if (collectCoverage) await page.coverage.startJSCoverage({resetOnNavigation: false}); await page.addInitScript(() => localStorage.setItem("loopx-pw-locale", "en")); - await page.goto(`${owner.base}/chat/?goalId=benchmark-goal&view=research`); + await page.goto(`${owner.base}/chat/?goalId=${capture.goal_id}&view=research`); const history = page.locator(".research-history"); await page.waitForFunction(() => document.querySelector(".research-history")?.getAttribute("aria-busy") === "false"); assert.match(await history.locator(".research-history-coverage").innerText(), new RegExp(`${capture.events.length} / ${capture.events.length}`)); @@ -99,17 +104,21 @@ export const researchRealBenchmarkScenario = { await absentDecisions.locator(":scope > summary").click(); assert.match(await absentDecisions.innerText(), /No explicit decision recorded/); await absentDecisions.locator(":scope > summary").click(); - await page.screenshot({path: resolve(outputDir, `research-real-grid-${arm}-desktop.png`), animations: "disabled"}); + await page.screenshot({path: resolve(outputDir, `research-real-${arm}-desktop.png`), animations: "disabled"}); await history.getByRole("button", {name: "Search research records", exact: true}).click(); const records = history.locator(".research-record-search"); const search = records.getByRole("searchbox", {name: "Search research records"}); - assert.equal(await records.locator(".research-record-entry").count(), capture.events.length); + assert.equal(await records.locator(".research-record-entry").count(), Math.min(50, capture.events.length)); + while (await records.getByRole("button", {name: /Show more records/}).count()) { + await records.getByRole("button", {name: /Show more records/}).click(); + } + assert.equal(await records.locator(".research-record-entry").count(), capture.events.length, "Later real records stay reachable across display pages"); const received = page.waitForEvent("download"); await records.getByRole("button", {name: "Download matches (JSON)", exact: true}).click(); const download = await received; const exported = JSON.parse(await readFile(await download.path(), "utf8")); - assert.equal(exported.goal_id, "benchmark-goal"); + assert.equal(exported.goal_id, capture.goal_id); assert.deepEqual(exported.coverage, {loaded_records: capture.events.length, exported_records: capture.events.length, history_complete: true, refresh_failed: false}); assert.deepEqual(exported.records.map(event => event.event_id), capture.events.map(event => event.event_id), "The full append order survives export"); for (const [index, original] of capture.events.entries()) { @@ -122,9 +131,10 @@ export const researchRealBenchmarkScenario = { // A source-authored 'confirmed' record can report no gain. Never infer // positive benefit from that label or rewrite the original conclusion. - const chosen = arm === "on" + const chosen = arm === "grid-on" ? capture.events.find(event => event.summary.includes("no measured gain from either tested family")) - : capture.events.at(-1); + : arm === "ad" ? capture.events.find(event => event.summary.includes("Not adopted. This is a scoped negative result")) + : capture.events.findLast(event => event.event_kind === "finding"); assert(chosen); await search.fill(chosen.event_id); const entry = records.locator(".research-record-entry"); @@ -135,13 +145,36 @@ export const researchRealBenchmarkScenario = { await entry.scrollIntoViewIfNeeded(); assert.equal(await page.evaluate(() => document.documentElement.scrollWidth > innerWidth), false); assert.equal(await entry.locator(":scope > summary > strong").innerText(), chosen.title); - await page.screenshot({path: resolve(outputDir, `research-real-grid-${arm}-record-${width}.png`), animations: "disabled"}); + await page.screenshot({path: resolve(outputDir, `research-real-${arm}-record-${width}.png`), animations: "disabled"}); } - if (arm === "on") assert.match(await entry.locator(".research-finding-sections").first().innerText(), /no measured gain/); + if (arm === "grid-on") assert.match(await entry.locator(".research-finding-sections").first().innerText(), /no measured gain/); + if (arm === "ad") assert.match(await entry.innerText(), /Not adopted/); const original = entry.locator(":scope > details").filter({has: page.locator("summary", {hasText: "Full original text"})}); await original.locator(":scope > summary").click(); assert.equal(await original.locator(":scope > p").first().innerText(), chosen.summary, "Long negative/unknown evidence is readable without abbreviation"); await original.locator(":scope > summary").click(); + if (arm === "ad") { + const first = capture.events[0]; + const versions = capture.events.filter(event => event.result_id === first.result_id); + assert.equal(versions.length, 39, "The real repeated question has 39 original updates"); + await search.fill(first.event_id); + await entry.locator(":scope > summary").click(); + const comparison = entry.locator(".research-record-comparison"); + await comparison.locator(":scope > summary").click(); + await comparison.getByRole("combobox", {name: "Compare with"}).selectOption(versions.at(-1).event_id); + await comparison.getByRole("checkbox", {name: "Include unchanged fields"}).check(); + for (const width of [1440, 390]) { + await page.setViewportSize({width, height: 900}); + await comparison.scrollIntoViewIfNeeded(); + await page.screenshot({path: resolve(outputDir, `research-real-ad-versions-${width}.png`), animations: "disabled"}); + assert.equal(await page.evaluate(() => document.documentElement.scrollWidth > innerWidth), false, "Real long version labels fit the reader"); + } + await comparison.getByRole("checkbox", {name: "Include unchanged fields"}).press("Escape"); + assert.equal(await comparison.evaluate(element => element.open), false); + await entry.locator(":scope > summary").evaluate(element => new Promise(resolve => { + requestAnimationFrame(() => resolve(element === document.activeElement)); + })).then(focused => assert.equal(focused, true)); + } await search.fill(""); await history.getByRole("button", {name: "Refresh history", exact: true}).click(); await page.waitForFunction(() => document.querySelector(".research-history")?.getAttribute("aria-busy") === "false"); diff --git a/tests/presentation/test_explore_history.py b/tests/presentation/test_explore_history.py index 23a36fccce..d4656c140d 100644 --- a/tests/presentation/test_explore_history.py +++ b/tests/presentation/test_explore_history.py @@ -22,6 +22,8 @@ ("edgebench-ad-checkpoint-history.json", "de04b679e9f382b650e891e5c7dd7146bc690500435c511f107c9d24da4f4efa", 145296), ("edgebench-grid-context-on-20261103-history.json", "292f493ee5d88e0f99546776ee2a46958e19517ceae192c10288276f270f404d", 80992), ("edgebench-grid-context-off-20261103-history.json", "d91047d06aa7ef43249bc79e70d8ba53d37a88240c124dcd17a75034ea524375", 20575), + ("edgebench-ad-20261023-history.json", "23a55562f8ad9b808a5633c479451b78b13ad5370a35190d04fab407e3824a72", 396701), + ("edgebench-schema-20261025-history.json", "5473191a1f23537a48f38ba82d0dd8ba8ed5ad7daffd944f28910e7d6e19e0ab", 271025), ]) def test_checkpoint_fixture_reconstructs_original_captured_log_bytes(name, digest, size): fixture = Path(__file__).resolve().parents[2] / "apps/presentation/dashboard/src/data/fixtures" / name