feat(sf): rewire last 5 skip-exceptions → dry (DriftDetection preflight + EvalJudge/Rationale/Replay/CF dry_run_llm)

[cipher813]

Closes the last keystone gap: the 5 documented shell-run skip-exceptions are flipped skip→dry. Under shell_run every substantive workload now boots + runs dry; ZERO skip-exceptions remain. All prerequisite dry flags were already MERGED on origin/main of their repos (data #261, research #202, backtester #225). LIVE Saturday SF — single-file SF-wiring change, precision-validated.

Per-state mechanism

State	Type	Mechanism under shell_run	Source PR
DriftDetection	spot	commands.$ States.Format($.preflight_args) → --preflight-only	data #261
EvalJudgeSubmitFirstSaturday	Lambda	Payload "dry_run_llm.$": "$.research_dry"	research #202
EvalJudgeSubmitWeekly	Lambda	Payload "dry_run_llm.$": "$.research_dry"	research #202
EvalJudgePoll	Lambda	Payload "dry_run_llm.$": "$.research_dry"	research #202
EvalJudgeProcess	Lambda	Payload "dry_run_llm.$": "$.research_dry"	research #202
RationaleClustering	Lambda	Payload "dry_run_llm.$": "$.research_dry"	research #202
ReplayConcordance	Lambda	Payload "dry_run_llm.$": "$.research_dry"	backtester #225
Counterfactual	Lambda	Payload "dry_run_llm.$": "$.research_dry"	backtester #225

Exact canonical dry var + why

$.research_dry — THE canonical shell-run LLM-dry signal. InitializeInput seeds it false on every run (absent path / the real Sat 02:00 PT firing unchanged); ApplyShellRunDefaults already sets it true under shell_run (it backed Research from the keystone). No new var invented: research #202 / backtester #225 PR bodies specify the dry_run_llm event flag, and reusing $.research_dry makes the absent-path guarantee automatic — the seed already exists, no extra InitializeInput plumbing required.

The change

• ApplyShellRunDefaults: removed skip_drift_detection / skip_eval_judge / skip_rationale_clustering / skip_replay_concordance / skip_counterfactual from the force-set JsonMerge blob → it now force-sets ZERO skip_*. Per-flag user overrides still win (merge order States.JsonMerge(defaults,$,false) unchanged). The Choice-gated CheckSkip<State> gates are LEFT INTACT (still valid for targeted operator skips — pinned by test_skip_gates_still_intact, all 16 retained).
• DriftDetection: literal commands array → commands.$ States.Array(...,States.Format('bash infrastructure/spot_drift_detection.sh{} 2>&1 | tee /var/log/drift-detection.log',$.preflight_args)), mirroring DataPhase1 exactly. {} immediately after the script token, no literal space; preflight_args carries its leading space inside the var → preflight_args="" reproduces the origin/main command char-for-char, " --preflight-only" yields exactly one separating space.
• 7 eval Lambda Payloads: added "dry_run_llm.$": "$.research_dry". EvalRollingMean was NOT touched — alpha-engine-research-eval-rolling-mean, no skip gate, never a keystone exception, pure historical-metric reader (out of scope).

Byte-identical proof (the non-negotiable invariant)

shell_run absent OR false ⇒ Saturday SF behaviorally byte-identical to today's real run:

• CheckShellRun.Default = CheckSkipMorningEnrich (unchanged). InitializeInput seeds preflight_args="", research_dry=false.
• Every spot States.Format resolves char-for-char to the frozen origin/main literal (verified via the test's own resolver for all 8 spot states, including DriftDetection now in the frozen fixture). Every eval Lambda dry_run_llm.$ resolves to false (handlers default it false ⇒ behaviourally identical to pre-rewire).
• tests/fixtures/sf_prekeystone_spot_commands.json regenerated via the established generator at preflight_args=""; now includes DriftDetection's pre-rewire origin/main literal command; the existing 7 entries are unchanged (git diff adds DriftDetection only).
• CI-safe: tests read only the committed fixture — no git show origin/main shell-out (that was the feat(sf): shell-run keystone — spot --preflight-only + Lambda --dry-run instead of skip #260 CI failure).

Test changes

• _SPOT_STATES → 8 (added DriftDetection); _DRY_LAMBDA_STATES → 11 (added the 7 eval states); _KEYSTONE_SKIP_EXCEPTIONS = set().
• test_shell_defaults_force_set_ZERO_skip_exceptions: blob force-sets no skip_*; none of the 16 workload skips (incl. the 5 ex-exceptions) appear.
• TestHappyPathTraversal: under shell_run skipped == set(); DriftDetection is VISITED (runs dry), not jumped past; absent shell_run = unchanged pre-keystone path.
• Module + class docstrings updated to the rewire semantics.

Validation

• JSON valid: 58 top-level states (91 incl. parallel branches).
• Full alpha-engine-data suite: 1351 passed, 1 skipped, 0 failed.

Zero skip-exceptions remain — every substantive task runs dry under shell_run (spots → --preflight-only, Lambdas → dry_run_llm).

🤖 Generated with Claude Code