verification-before-completion

Verify code quality before committing by running fresh test, build, and lint commands, capturing full output, and checking exit codes to enforce evidence-backed completion claims.

Preventing false completion claims in CI workflows requires a strict Identify, Run, Read, Verify, Claim sequence that validates success assertions using fresh command execution and full output inspection.

Validate agent-reported build successes by re-running verification commands, inspecting exit codes and full outputs, and requiring explicit failure counts before accepting completion status.

This verification approach works with existing pre-commit hooks, linters, and regression checks by acting as a gate that enforces fresh execution and evidence collection before declaring work ready.

Evidence-backed status reporting for pull requests reduces regressions and rework by requiring captured command output and explicit exit codes rather than accepting unsupported claims of passing tests or builds.

Relying solely on exit codes for regression checks misses hidden failures, so this approach mandates full output inspection alongside exit code checking to ensure genuine verification.