Repository navigation
fix: drop the re-run helper's pull-requests permission (callers fail at startup) - #66
Conversation
…ller at startup A called workflow can't request more than the calling job grants. The callers grant contents: read + actions: write, so pull-requests: read on rerun-losing-run made every run a startup_failure (zero jobs, the required check never reports). Read the PR's current head from the run's own pull_requests field instead, which needs no extra permission. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
|
Note Some coding-guideline sources were skipped; the reason is listed next to each one. Other guidelines were applied as usual. On a public repository, sources from your configuration are listed without their names. Skipped guideline sources (1)No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configuration
📒 Files selected for processing (1)
Included review availability: This review used your included allowance. 3 included reviews remain after this review. Your included PR review attempts over the past 7 days set your current allowance at 5 reviews per hour. 📜 Recent review details🔇 Additional comments (2)
WalkthroughThe workflow removes pull-request read permission from ChangesInline test rerun validation
Priority: ➖ Normal Estimated code review effort: 2 (Simple) | ~10 minutes Merge Risk: ⚪ Minimal · up to This change fixes the startup failure caused by the excess permission request. The remaining uncertainty is whether the embedded PR head stays current after a new push. It is worth confirming on the next LU PR run after merge. 🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
Comment |
…nd this week's lessons Drops examples that pointed at check_auth.yml, the old run_tests.yml, refresh_token.yml and be_market's fork. Corrects two claims measured wrong in lu_market: a re-run also reads the current secret, and a dispatched retry never clears a PR's red check (only a re-run does). Adds checklist 18: a job-level permission in a reusable workflow beyond what a caller grants fails that caller's every run at startup (#66). Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…#68) * chore: remove the old run_tests.yml, check_auth.yml and update_templates_production.yml All four markets now call run_tests_inline_auth.yml, so nothing calls the old run_tests.yml or check_auth.yml any more (checked every workflow in be/nl/lu/uk, plus an org-wide code search). update_templates_production.yml only ran on this repo's own PRs, called the old run_tests.yml and failed on every run. README: drop the two old sections, describe run_tests_inline_auth.yml on its own. Comments that pointed at the removed files are updated. update_templates_review.yml and add_shared_parts_*.yml stay for now: a colleague may reuse ideas from them for a deployment pipeline. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> * chore: also remove update_templates_review.yml and add_shared_parts_*.yml Dormant label-triggered workflows that only ran on this repo's own PRs and predate the current login handling. push_to_review_firm.yml is the maintained successor of update_templates_review.yml. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> * docs(skill): update writing-bso-workflows for the removed workflows and this week's lessons Drops examples that pointed at check_auth.yml, the old run_tests.yml, refresh_token.yml and be_market's fork. Corrects two claims measured wrong in lu_market: a re-run also reads the current secret, and a dispatched retry never clears a PR's red check (only a re-run does). Adds checklist 18: a job-level permission in a reusable workflow beyond what a caller grants fails that caller's every run at startup (#66). Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> * docs: fix stale sampler references (check_auth cron, REPO_ACCESS_TOKEN scopes) Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> * docs: point remaining run_tests.yml references at run_tests_inline_auth.yml; sampler token needs Pull requests write Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Description
Urgent: every LU liquid-test run fails at startup since #63 merged (e.g. lu_market run 37429525996:
startup_failure, zero jobs). The required check never reports, so LU PRs are blocked. nl/be would hit the same as soon as they switch.Cause: #63 gave
rerun-losing-runpull-requests: read. A called workflow can't request more than the calling job grants, and callers grantcontents: read+actions: write, so GitHub rejects the whole run.Fix: drop that permission. The PR-closed/moved check now reads the PR's current head from the losing run's own
pull_requestsfield, which the runs API already returns. It lists only open PRs, with their live head (checked on real runs: open PR → current head, merged PR → empty), so no extra API call is needed.Type of change
Test plan
actionlint: no new findings🤖 Generated with Claude Code