This is the formal review handoff for the aegis method pack. It spins up a fresh code-reviewer subagent with zero context from your session, forcing you to articulate exactly what you built, what spec it should match, what evidence already exists, and what baseline or compatibility boundary matters. You're required to answer nine questions before dispatch, including what happens to any old logic that should retire. The reviewer checks for bugs and risks first, then tests and evidence gaps, then baseline drift and missing ADRs. It's mandatory after each subagent task and before any high-risk merge. The review is advisory only, not authoritative completion, but it catches compound errors early when fixing them is still cheap.
npx -y skills add ganyuanran/aegis --skill requesting-code-review --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
github/awesome-copilot
alirezarezvani/claude-skills
microsoft/win-dev-skills
github/awesome-copilot