verification-before-completion

Fresh verification requires running tests, lint, type checks, and CI checks, then reading their outputs to validate completion status. This workflow ensures claims are backed by concrete evidence, preventing premature success declarations during code changes or PR reviews.

Fresh verification evidence is the validated output from recently executed tests, lint, and type checks. It serves as a quality gate to determine the truth of a completion status before any success claim is made during code changes, PR reviews, or deployment gates.

To stop premature completion claims in PR reviews, apply a guardrail workflow that identifies verification commands, executes the full suite, reads outputs, and declares success only after verifying explicit evidence. This ensures all status updates are auditable and non-deceptive.

Yes, full verification requires running tests, lint, type checks, and CI checks together. Evaluating the combined outputs ensures comprehensive evidence reporting and accurate completion status determination before making any success claims about code changes.

Relying solely on CI checks limits completion claims if fresh local verification is skipped. The guardrail workflow requires executing and reading outputs from tests, lint, and type checks directly to ensure safe, auditable evidence rather than assuming prior CI status remains valid.