verification-before-completion

Verification-before-completion enforces mandatory testing and build checks before declaring tasks complete. Run your verification commands—tests, linters, builds—capture their output, and only assert success after confirming the results meet your criteria, eliminating false completion claims.

Execute your test suite and linter commands fresh, review the command output directly, and require explicit confirmation that results succeeded before creating a commit or pull request. This evidence-based approach replaces assumptions with verifiable proof.

Yes. Specify verification commands (like `cargo test` or `npm test`), integrate them as gates before task or epic completion, and use summary parsing to keep context efficient. This prevents shipping broken features by making verification a mandatory step.

Fresh verification captures the actual state of current changes. Earlier results may not reflect recent code modifications. Running commands again and analyzing their output ensures you have concrete evidence that the work truly meets success criteria.

Failed verification output blocks completion claims. You then fix the issues, re-run the commands, and only declare work complete once the fresh output confirms success. This gate prevents incomplete work from advancing to commit or deployment.