verification-before-completion

Verification-before-completion enforces running fresh test commands, capturing full output and exit codes, and presenting that evidence before declaring tests pass. This prevents false completion claims by requiring concrete proof rather than assumptions.

Execute the verification commands fresh, analyze the complete output logs and exit codes, then report results with that evidence attached. This evidence-first approach confirms or refutes claims and reduces rework from unverified assertions.

Yes. The Skill provides a repeatable checklist process—Identify, Run, Read, Verify—that applies to any environment claiming task completion: tests passing, builds succeeding, lints clean, or regressions verified across pipelines and local workflows.

Fresh verification captures current state and prevents stale claims. Running commands again, reading full output, and checking exit codes ensures the evidence reflects actual present conditions, reducing miscommunication and false positives.

By enforcing evidence before any completion claim, verification gates catch unproven assertions early. This guardrail reduces rework, prevents propagation of false positives into downstream tasks, and improves team confidence in reported results.