Key-presence refusal at every gate: a pinned model is not enough

Owner-diagnosed: the turn gate checked only model coordinates, so a
pinned slot whose provider key was deleted passed the gate and died
mid-stream masked to 'An error occurred.' The key is now gated where
the model is — by what would RUN (env-first keyValue truth, so an
env-shadowed key passes).

- core/settings.ts: the provider_key_missing envelope family (code,
  provider, slots, human line naming the provider from AUTH_PROVIDER_KEYS
  labels) + ProviderKeyMissingError + isProviderKeyRefusal matcher.
- chat-handler turn gate: refuses 503 before beginTurn when the resolved
  conversation's provider key is gone (slots: [strong]).
- Job runner: same check via an injected dep (agent layer cannot read
  the key store) — a keyless engine refuses the spawn through the
  existing failure-shape convention, no phantom run.
- Extraction pipeline: refuseIfStagesUnready refuses BEFORE any batch
  (first offender: extractor before auditor); the trigger skips info-
  level, /api/graph/import surfaces the 503 envelope per file.
- Tests: 8 new pins (turn gate both ways, runner refusal, pipeline
  first-offender + pass-through, envelope factory, matcher); the three
  turn-driving suites seed OPENROUTER_API_KEY like any real setup.
- Doc fix: stale EXTRACTION.chunkTokens comment in the manual-extract
  route (left over from eed4b4e).

Known residual (pre-existing, flagged for master): the manual extract
route logs BOTH config refusals as console.error (fire-and-forget path);
the coverage ledger keeps everything uncovered either way.
This commit is contained in:
amosblomqvist
2026-09-24 22:51:12 +03:00
parent 085fdc1a98
commit e4049e4bf8
13 changed files with 411 additions and 36 deletions
+2 -1
View File
@@ -3,7 +3,8 @@
* (vision-extraction-coverage.md): run the pipeline over ONE session's
* uncovered content, whatever its size — the user's answer to an aborted
* run, a pipeline error, or a session that went dormant under the chunk
* budget. The pipeline itself batches under EXTRACTION.chunkTokens and
* budget. The pipeline itself batches under the graph.extractionChunkTokens setting (extraction-UX redesign,
* 2026-09-24) and
* checkpoints coverage per merged batch, so this is the same run the
* trigger fires, just user-initiated. 409 while a run is already in
* flight (the dock is the source of truth for that); failures are logged