React Doctor — A Diagnostic Tool That Validates AI-Generated React Code Through Static Analysis

React Doctor, a diagnostic tool for validating the reliability of React code written by AI coding agents, has been released. Running the npx react-doctor@latest command to scan a project provides diagnostic results covering state management, side effects, performance, security, accessibility, and overall architecture, along with a score between 0 and 100.

Key Features

• AI Agent Workflow Support: Skills can be automatically registered with major agents such as Claude Code, Cursor, Codex, and OpenCode. This supports a complementary structure where one agent writes code and another agent fixes issues based on the diagnostic results.
• High-Speed Static Analysis: Built on oxlint with over 100 built-in rules, enabling very fast analysis without a separate ESLint runtime. It offers various framework-specific presets for Next.js, TanStack, React Native, and more.
• React Compiler Optimization: In environments where React Compiler is enabled, it immediately detects unnecessarily remaining manual memoization code.
• CI/CD and Automation: Natively supports GitHub Actions, leaving diagnostic results as inline annotations and comments on each PR. The --diff mode allows scanning only changed files to reduce CI time.
• Extensibility and Programmatic Access: Uses an Effect v4-based streaming pipeline to run linting and dead code analysis in parallel, and diagnostic results can be processed directly in Node.js code via @react-doctor/api.

While existing ESLint plugins have focused on configuring individual rules, React Doctor positions itself as an 'Audit Tool' that checks overall code quality immediately upon installation, aiming to solve the code quality management problem in AI-driven development environments.