[regression] Daily CI Regression Audit — 32208989497

[github-actions]

Summary

CI failure rate (last 20 completed runs on .github/workflows/ci.yml): 3/20 (15%) — above the 10% healthy threshold. No consecutive failure streak (failures interspersed with successes). Recommended action: investigate the lock-file drift regression below; it appears to be a real, recurring issue rather than flakiness.

Failure Details

Top failure patterns (3 failed runs inspected):

1. Lock file drift — smoke-cursor.lock.yml (2 of 3 failures, runs 31186311702 and 31185851384, both on main, ~5 min apart on 2026-08-07)

   • Job "Check workflow markdown/lock drift" fails with:

     ERROR: Workflow lock files are out of sync with their markdown sources.
     The following lock files differ from what would be generated by compilation:
       .github/workflows/smoke-cursor.lock.yml
     

   • Identical error in both runs — this is a consistent regression, not flaky. Indicates a commit landed on main that modified smoke-cursor.md (or shared compiler logic affecting it) without regenerating/committing the corresponding .lock.yml via make recompile.

2. health-smoke-copilot job — Go module fetch failure (1 of 3 failures, run 31183472811)

   • Build step fails during go build:

     ##[error].../go-sdk@v1.7.0/internal/json/json.go:13:2: github.com/segmentio/encoding@v0.5.4:
     reading (proxy.golang.org/redacted) 403 Forbidden
     make: *** [Makefile:42: build] Error 1
     

   • This looks like a transient upstream Go module proxy issue (403 Forbidden from proxy.golang.org), not a code regression — likely flaky/network-related.

Open regression/bug issues: No open issues found with labels bug + regression on github/gh-aw.

Generated by 🧭 Daily Regression Audit · auto · 70.3 AIC · ⌖ 7.2 AIC · ⊞ 6.9K · ◷

• expires on Aug 20, 2026, 6:41 PM UTC-08:00