verify-before-claim

To prevent false completion claims in code review, you must require fresh verification evidence before stating work is done. This entails running a current proof command, capturing the full output, checking pass/fail counts, and reporting residual risk before stating a pull request is passing.

Fresh verification evidence for a bug fix or build check is observable proof generated immediately before a success claim. It includes executing a proof command, inspecting actual output, verifying the exit code, and documenting failure counts to prove the status is not overstated.

Agents say tests are passing without checking because they lack a verification gate requiring current proof. Implementing a rule that forces an inspection of fresh test output, exit codes, and failure counts before any completion statement stops unsupported claims and ensures accurate status reporting.

To verify code changes before committing or updating documentation, run the freshest proof command available for the claim. Inspect the complete observable output, validate the exit code, check pass and failure counts, and report actual status with any residual risk before declaring completion.

Yes, requiring proof commands works for documentation updates and status checks by applying a fresh verification gate to any statement of work done. It forces capturing observable evidence, exit codes, and residual risk reporting before allowing claims that documentation or status checks are complete.

The limitation of gating completion claims on verification evidence is that it requires a valid, current proof command to exist for every task. If a task lacks an observable inspection command or measurable exit code, generating the required evidence to prove residual risk becomes difficult.