Skip to main content
Lab Grimoire
TW EN
☕ Coffee
Spellbook / react-doctor: Zero-Config React Code Quality Scanner with Agent Skill Support
software

react-doctor: Zero-Config React Code Quality Scanner with Agent Skill Support

A deterministic CLI that scans your React codebase for state management, performance, security, and accessibility issues. Best for teams using AI coding agents alongside React. Free, zero-install via npx, with built-in GitHub Actions integration. Telemetry is on by default but can be disabled.

Best For

React developers who want automated, deterministic code quality checks across state management, effects, performance, security, and accessibility. Especially valuable for teams where AI coding agents generate React code and a systematic second pass is needed before merge.

How I Actually Use It

I run npx react-doctor@latest on my React + Firebase project before committing. The tool produces a categorized audit report that catches patterns my AI coding agent sometimes gets wrong: unnecessary re-renders, effect dependency array issues, and missing accessibility attributes.

For CI, I set up the GitHub Actions integration so that pull requests only flag newly introduced issues. This keeps the noise down and lets the team address tech debt incrementally rather than being overwhelmed by a wall of warnings on day one.

I also installed it as a coding agent skill via npx react-doctor@latest install, which lets Claude Code and Cursor run the scan as part of their own review loop.

Where It Is Strong

  • Zero installation barrier: npx react-doctor@latest runs immediately with no setup
  • Deterministic scanning: results are stable and reproducible, not LLM-based guesswork
  • Broad coverage: state, effects, performance, architecture, security, and accessibility in one pass
  • Framework-aware: supports Next.js, Vite, TanStack, React Native, and Expo out of the box
  • Agent-native: built-in skill installation for coding agents like Claude Code and Cursor
  • CI-friendly: GitHub Actions integration reports only new issues per PR

Where It Fails

  • Telemetry is enabled by default; you must add --no-telemetry to opt out
  • Rule customization requires manual doctor.config.ts configuration with limited documentation on advanced tuning
  • React-only: if your project is not in the React ecosystem (Vue, Svelte, Astro-only), this tool has nothing for you

Pricing, Difficulty, and Risk

Price: Free and open-source under MIT license.

Difficulty: Beginner-friendly. A single npx command gets you a full audit report. No configuration required for the default scan.

Risk: Low. As a static analysis tool, it never executes the code it inspects. Telemetry is the main concern: it is on by default and collects execution metadata (the maintainers state no source code is collected). Disable with --no-telemetry if this matters to your organization.

Verdict

If you ship React code and want a fast, deterministic quality gate that also plays nicely with AI coding agents, adopt react-doctor. Skip it if your stack is not React-based.

Source