Verification Before Completion

Verification before completion requires running the specific test, build, or quality command that confirms the fix works, inspecting the full output and exit code, and only then declaring success. This eliminates false positives by anchoring completion claims to concrete evidence rather than assumptions.

Execute the complete verification command—linting, tests, or builds—in full, read the output carefully to confirm all checks pass, and verify the exit code indicates success. Only after confirming evidence should you claim readiness or create the PR.

Incomplete verification allows premature completion claims. Without running and inspecting verification commands end-to-end, you miss failures that only surface during actual testing, builds, or code reviews, creating rework and trust issues downstream.

No. Verification before completion enforces that all completion claims—whether for bug fixes, features, or code reviews—must be gated on evidence from the actual verification command. Partial or skipped verification defeats the process.

The verification process identifies common rationalizations and red flags that lead to skipped steps. By following the strict gate-function process—identify command, run it fully, read output, confirm exit code—you enforce evidence-based claims and eliminate excuses.

Inspect the full output and exit code to understand what failed. Do not claim completion until the verification command succeeds completely. Address the failures, re-run verification, and confirm success before proceeding.