React Doctor is an open-source scanner for React projects that helps developers catch weak patterns before they ship, with a special focus on code written by coding agents. The tool runs from the command line with npx react-doctor@latest, inspects a React codebase, and returns a 0 to 100 health score plus diagnostics. The README describes checks across state and effects, performance, architecture, security, accessibility, and framework-specific React patterns.
The project is built for modern React teams rather than generic static analysis only. It supports Next.js, Vite, React Native, mixed monorepos, GitHub Actions, CLI usage, Node.js API usage, oxlint, and an ESLint plugin path. Rules are toggled based on detected framework and React version, which helps reduce irrelevant advice. For pull requests, React Doctor can post sticky comments, emit annotations, expose a score output, and fail builds based on diagnostic severity or a team-defined score floor.
The agent workflow is the sharpest use case. React Doctor can install best-practice instructions for coding agents by writing files such as SKILL.md, AGENTS.md, or .cursorrules. That turns it into both a reviewer and a guardrail: agents get guidance before writing code, and the scanner checks the resulting project afterward. Teams using Claude Code, Cursor, Codex, OpenCode, Windsurf, or other agents can add it to local checks or CI to catch regressions early.
React Doctor is not a substitute for human review or product-specific testing. A score is a signal, not a final judgment, and new rules can lower scores over time. The README recommends pinning versions when enforcing score floors because rule changes can affect CI. It also notes that dead-code detection through knip was removed in v0.2, so teams that need that check should run knip separately.
OpenTools lists React Doctor as a developer AI tool because it targets a new real problem: agent-written React that compiles but is messy, slow, inaccessible, or hard to maintain. It gives teams a repeatable way to score, review, and gate React quality as AI coding tools produce more of the first draft.
For evaluation, treat React Doctor as a builder-focused open-source project rather than a managed SaaS. Review the upstream README, license, install path, and issue activity before adopting it. Teams should test it in a disposable repository or development environment first, document the exact version they use, and keep production workflows behind normal code review, monitoring, and rollback practices.
React Doctor works best when it becomes part of the normal React delivery loop. Run it locally before large agent-generated changes, add the GitHub Action for pull requests, and choose a score floor that matches the team’s tolerance for warnings. Because rule sets evolve, pin versions in CI and review new findings before turning stricter checks into merge blockers.