daily-fix: Step 10d lint verdict consumed before gh pr merge
kind: infra#wf-fix#wf-fix-fp:72ecdd823040#daily-auto-filed
track:Highlight text on any card (body, plan, or an event) to anchor a comment, or leave a whole-task comment here. Mention @claude to summon a reply.
No comments yet.
Overview / Motivation
Auto-filed by /daily 2026-07-06 (route 2: behavior/logic change -> independent review), from the nightly transcript problem sweep.
Goal
Consume (rm) the lint-verdict file only AFTER the gh pr merge attempt returns success; invalidate the verdict on new lint-surface commits instead of on read, so a rebase-refusal→squash retry stays certified by the still-valid gate run rather than a hand-written verdict.
Workflow gap
- Bug observed: The Step 10d merge template rm -f's the consume-once lint-verdict file BEFORE
gh pr mergeruns; when the merge fails for non-lint transport reasons ('can't be rebased' / merge conflicts) the PASS verdict is orphaned and the retry has no gate. #1082's session hand-recreated the gate file (echo pass > /tmp/issue-1082-lint-verdict.txt) — a self-attested gate write, exactly what consume-once was meant to prevent. Hit twice on 2026-07-06: sessions 24257d06 (#1081, PR #842) and d7ae9a3b (#1082, PR #843), both the documented #1041 rebase-refusal→squash shape. - Why it is a workflow gap: the failure originates in the workflow surface / shared helper named below, not in any one experiment.
- Confidence (emitter): medium
Scope / surfaces
- Primary target:
.claude/skills/issue/SKILL.md - Grep the workflow surface for the pattern before editing and update every hit; list them in the plan.
Constraints / invariants
scripts/workflow_lint.py --check-asks+--check-referencesstay green; ruff clean on touched files; relevant tests pass.
Provenance
- workflow_fix_target: .claude/skills/issue/SKILL.md
- source: /daily 2026-07-06 problem sweep (transcript-mined)
Activity