stacklok / toolhive-studio / 24721685183

70%

push

github

feat(ci): add automated bug-fix agent and triage cron (#1938)

* feat(ci): add TDD bug-fix agent and triage cron

Add automated bug fixing via TDD approach:

- Bug-fix agent workflow triggered by `auto-fix` label on Bug issues
- Two-phase TDD: Opus writes failing test (red), Sonnet applies fix (green)
- Up to 3 attempts per issue with 5 internal retries per phase
- Hard gates verify test fails before fix and all checks pass after
- Daily triage cron (Sonnet) evaluates open Bug issues for auto-fix suitability
- Skips issues with comments, cross-repo fixes, and existing PRs
- Only comments on issue after all 3 attempts are exhausted

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* style: fix prettier formatting in bug-fix-tdd skill files

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* feat(ci): add direct fix fallback when TDD cannot reproduce bug

- Add Phase 2b: when Phase 1 writes a test but it passes (bug not
  reproduced), fall back to a direct fix without regression test
- Reduce Phase 1 internal retries from 5 to 3
- Increase Phase 2/2b max-turns from 80 to 150
- Rename job from "TDD Bug Fix" to "Bug Fix"
- Update skill docs with Phase 2b documentation

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* fix(ci): address Copilot review feedback on bug-fix workflows

- Fix grep exit code killing script under bash -e (|| true)
- Add Write to Phase 2 allowedTools (needed for pr-body.md)
- Add Write to triage allowedTools (needed for triage-results.json)
- Fix .comments jq filter (number, not array)
- Align change detection with git add (merged into single step)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* fix(ci): rename step and fix git clean glob pattern

- Rename "TDD Bug Fix" step to "Bug Fix"
- Fix git clean glob: ** requires globstar, use plain -fd instead

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* fix(ci): remove auto... (continued)

3844 of 6337 branches covered (60.66%)

5983 of 9023 relevant lines covered (66.31%)

119.48 hits per line