AI Code Review Toolkit

AI coding agents write code that looks finished but fails differently than human code: dropped auth checks, tests that mock the function they test, hallucinated API calls. This toolkit gives you the 6 recurring failure patterns, a 45-item review checklist, a risk-scoring rubric, and a PR tracker spreadsheet to see which AI tool causes the most rework. For solo devs and small teams using Claude Code, Cursor, or Copilot.