Deploy: failure
workflow_run_id: 25341303587
url: https://github.com/justnau1020/finlet/actions/runs/25341303587
head_sha: 214546c876a8cfafa178d507c362d98ef05994ee
finished_at_utc: 2026-05-04T20:43:37Z

Job outcomes:
  test (3.12)  : success
  test (3.13)  : success
  e2e          : failure
  deploy       : skipped (e2e gate)

Failing test:
  tests/e2e/journey-16-benchmark-blindness.spec.ts:49
  "benchmark API responses intercepted for blindness"
  Expected: API response must not contain scenario name "bull_run"
  Received: response body includes "bull_run"
  (139 passed, 1 failed, 5 skipped in journey-16 file)

Attribution:
  Iteration scope was leaderboard component metrics (Team A) + ledger maintenance (E) + docs (D).
  Zero changes touched benchmark/, /api/benchmark/*, or scenario name redaction logic.
  Last successful Deploy Finlet run on main was 9c3843e (12h prior); journey-16 was not exercised between.

Classification per bug-hunt skill rules:
  Deploy outcome: failure → retest SKIPPED.
  Patch-ineffective signal: NO — patches did not reach prod, which is the precondition for "patch ineffective."
  No pending_recurrences.jsonl entry warranted from this deploy outcome.

Follow-up (separate from this iteration):
  Investigate journey-16 — likely either:
    (a) A real benchmark-blindness leak introduced by an earlier commit in 17f1f2d..214546c (low likelihood — none of those commits touch benchmark code), or
    (b) A flaky test (most likely — only one occurrence; same scenario-name test passed on prior deploys).
  Owner: api/benchmark team. Filing as a separate work item.
