checklist-runner

You can automate markdown checklist validation by resolving a checklist name, loading local evidence like code and test results, and scoring each item as PASS, FAIL, PARTIAL, or N/A. The execution produces a computed pass-rate summary with final decision recommendations.

Running a pre-push quality gate using markdown is best achieved by executing the checklist in YOLO mode. This autonomously evaluates all sections against local evidence like docs and git diff, generating a final pass/fail decision without requiring interactive human confirmation.

Yes, you can pause checklist execution for human confirmation by using the interactive execution mode. This mode pauses between sections to allow human review of the aggregated evidence before proceeding to the next validation checkpoint.

Checklist validation does work with fuzzy name matching when the exact file is missing. It searches expected development paths to find the requested checklist and suggests matching alternatives when an exact name is not found.

Evidence sources used to score QA validation items include relevant local docs, code files, test results, and the current git diff. The system aggregates this evidence to determine if each markdown checklist item satisfies story, PR, or release requirements.