A structured code review system that runs three parallel adversarial reviewers (Blind Hunter, Edge Case Hunter, Acceptance Auditor) then triages their findings into actionable categories. Trigger it with "run code review" and it walks through a step-by-step workflow: gathering context from your project config, launching the parallel reviews, and organizing results. Built on bmad's step-file architecture, which means it loads one instruction file at a time and waits for you at checkpoints instead of racing ahead. The adversarial angle is deliberate: each reviewer attacks from a different perspective to catch what a single pass would miss. Good for teams who want consistent, thorough reviews without the ceremony of traditional checklists.
npx skills add https://github.com/bmad-code-org/bmad-method --skill bmad-code-review