verification-before-completion

Verify test evidence before claiming a pull request is ready by mapping the completion claim to a specific verification command, executing it fresh, and interpreting the full output and exit status before reporting success. This prevents premature or false completion claims.

Evidence-based verification for release readiness is the discipline of requiring fresh test, lint, or build output to back up passing-status claims. It forces you to select a narrow verification command, run it, and only state the claim with supporting evidence or report blockers.

Validate bug fixes before pushing code by selecting an evidence-matching verification step, running the full command to obtain current output and exit status, and only then stating the claim with supporting evidence. This avoids assumptions or partial checks.

When verification commands cannot run for a delegated task, you should state the blockers and use alternative evidence types like inspection or reproduction steps. This maintains claim-to-evidence discipline without skipping the verification process entirely.

Prevent false success claims in CI/CD pipelines by enforcing guardrails against dishonesty and partial checks. Require mapping assertions to specific verification commands, executing them fresh, and interpreting full output and exit status before reporting completion.

Yes, evidence-based verification works for delegated-agent handoffs by applying the same claim-to-evidence discipline. Delegated agents must execute fresh verification commands and provide current output and exit status before claiming task completion.