react-doctor

Why we included this project

The recurring problem with AI-generated React isn't that it fails outright; the code usually runs, but it quietly accumulates broken effects, needless re-renders, and accessibility gaps that the model keeps repeating. React Doctor scans a codebase deterministically and reports issues in state and effects, performance, architecture, security, and accessibility, giving you reproducible findings instead of a model's opinion. It runs across the usual React setups, including Next.js, Vite, and React Native. The agent handoff is the genuinely useful part: install it as a skill for Claude Code, Cursor, Codex, or OpenCode, and the CLI passes scan results to the agent's terminal so it can triage and fix them in a scan-filter-fix-validate loop. It also runs in CI, where it reviews pull requests and flags only the issues your change introduced, so you can gate merges on a clean diff rather than a growing backlog.