What problem does it solve?
This Skill solves the pervasive issue of developers and AI agents making unsubstantiated claims that work is complete, tests pass, bugs are fixed, or builds succeed without running actual verification. This behavior leads to broken code being shipped, wasted time on rework, and broken trust between collaborators and AI assistants.
Core Features & Use Cases
- Mandatory Verification Gate: Enforces a strict 5-step process that requires running fresh, full verification commands, checking exit codes and full output, and providing evidence before any completion or success claims are made.
- Anti-Rationalization Guardrails: Lists common excuses for skipping verification (e.g., "I'm confident", "Just this once") and their rebuttals, plus clear red flags to stop invalid claims before they happen.
- Use Case: Apply this Skill before committing code, creating pull requests, delegating tasks to AI agents, or claiming a bug is fixed to ensure only verified, working work is marked as complete.
Quick Start
Use the verification-before-completion skill to confirm all tests pass and the build exits successfully before you claim your recent code changes are ready to commit.