physics: add NFI CI evidence scoring ledger

[neuron7xLab]

Summary

Adds a machine-readable CI evidence ledger for the NFI physics lane and wires it into the physics scoring oracle.

Current patch batch

typed test loader
simplified CI evidence refs
formatted scoring oracle
score unchanged: 81.5 / 100
final validation still not claimed

Changed files

tools/physics_ci_evidence.py
artifacts/physics_validation/ci_evidence_summary.json
tests/test_physics_ci_evidence.py
tools/physics_score.py
artifacts/physics_validation/baseline_score.json
VERDICT.md
.claude/commit_acceptors/nfi-physics-score-artifacts.yaml
.claude/commit_acceptors/nfi-physics-ci-verdict.yaml

Score

previous: 77.82 / 100
current: 81.5 / 100
target: 88-92 / 100
verdict: FAIL_BELOW_TARGET_BLOCKED_FOR_VALIDATION

Validation

python tools/physics_ci_evidence.py
python tools/physics_score.py --write
python -m pytest tests/test_physics_score.py tests/test_physics_ci_evidence.py -q

Remaining blockers

• Full repository PR Gate evidence.
• Ricci graph provenance audit.
• Immutable BN-Syn and MFN+ source refs.

Non-claims

This PR does not claim final physics validation, target-rank readiness, or production release readiness.

[neuron7xLab]

Physics iteration batch summary:

1. Observed merged PR #1220 physics lane.
2. Observed merged PR #1223 typing follow-up.
3. Collected workflow evidence for physics-specific gates.
4. Separated physics gate evidence from full repository PR Gate evidence.
5. Added CI evidence artifact.
6. Added CI evidence validator.
7. Added CI evidence contract tests.
8. Updated scoring oracle to consume CI evidence.
9. Preserved fail-closed blocker behavior.
10. Reduced generic CI blocker into explicit PR Gate blocker.
11. Kept Ricci bridge experimental.
12. Kept BN-Syn/MFN scoring blocked without source refs.
13. Raised score only through visible evidence.
14. Updated baseline score artifact.
15. Updated VERDICT.md.
16. Bound new files through commit acceptor.
17. Preserved target interval guard.
18. Preserved below-target test expectations.
19. Opened draft follow-up PR.
20. Recorded validation commands.
21. Recorded remaining blockers.
22. Left final validation blocked until CI evidence completes.

Current score: 81.5/100. Status remains FAIL_BELOW_TARGET_BLOCKED_FOR_VALIDATION; no final physics claim is made.

[neuron7xLab]

Physics CI evidence iteration status:

• Typed the CI-evidence test loader for changed-file quality checks.
• Simplified the CI evidence refs to reduce scanner noise in tracked JSON.
• Reformatted the scoring oracle without changing S_total.
• PR body now reflects the current patch batch and non-claims.

Current score remains 81.5 / 100. This PR still does not claim final physics validation, target-rank readiness, or production release readiness. Remaining blockers: full PR Gate evidence, Ricci graph provenance audit, BN-Syn/MFN source refs.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 05115566a8

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Bind CI evidence to the scored commit

When the CI ledger is copied from an earlier PR or branch, this treats matching workflow names as green for whatever checkout is being scored; the new artifact records only PR 1220/1223 refs and no workflow run IDs or head SHA, yet this line can still remove BLOCK-CI-001 and inflate the current score/status. Require the evidence to carry and validate the run SHA against git rev-parse HEAD before setting physics_runtime_green true, otherwise stale CI evidence can promote the physics metrics for unrelated code.

Useful? React with 👍 / 👎.