Quality Playbook

This is a forensic quality audit system that runs a multi-phase investigation on any codebase, starting with domain-driven exploration rather than mechanical test generation. It derives behavioral requirements from what it finds, then runs three-pass code reviews, multi-model spec audits (the "Council of Three"), and produces regression tests with TDD-verified patches. The author claims it catches 35% of defects that structural review misses by grounding everything in fitness-to-purpose analysis instead of coverage metrics. The artifact is substantial, over 1,500 lines with phase-by-phase prompts and a CLI orchestrator. Worth trying if you're tired of test generators that pass CI but ship bugs, though the setup cost is real.

Install to Claude Code

npx -y skills add github/awesome-copilot --skill quality-playbook --agent claude-code