Your agent writes bad React. This catches it
13.5k
Stars
431
Forks
58
Open issues
17
Contributors
AI Analysis
React Doctor is a deterministic linter and code review tool that scans React codebases to identify issues in state management, effects, performance, architecture, security, and accessibility. It serves developers and AI coding agents (Claude Code, Cursor, etc.) who want automated detection of React anti-patterns, and works across all React frameworks including Next.js, Vite, React Native, and Expo. This is a specialized tool for React developers and agent developers, not for general-purpose c...
Inferred from signals mentioned in the README (tests, CI, type safety) — not a review of the actual code.
AI's overall editorial judgment — not an average of the bars above, can weigh other factors too.
React Doctor audits AI-generated React code for bugs, anti-patterns, and performance issues via CLI and CI
React Doctor is a deterministic static-analysis CLI tool that scans React codebases for problems across state management, effects, performance, architecture, security, and accessibility. It is explicitly positioned for teams using AI coding agents (Claude Code, Cursor, Codex, etc.) that tend to produce suboptimal React patterns. It runs as a one-off npx audit, as an installable 'skill' agents can learn from, or as a GitHub Actions CI check that isolates only new issues. It targets React developers across all major frameworks and any team adopting AI-assisted development workflows.
Created in February 2026 by millionco (the team behind the Million.js performance library), React Doctor appears to be a deliberate response to the rapid proliferation of AI coding agents generating React code without idiomatic correctness guarantees.
Accumulated 13,070 stars in roughly four months, with 215 stars in the last 7 days — indicating sustained organic interest rather than a single viral spike. Growth is plausibly driven by the AI coding agent narrative, which is a high-traffic topic in early 2026, combined with credibility from millionco's prior React ecosystem work (Million.js). The framing 'your agent writes bad React' is precise enough to resonate with a real and growing pain point.
NPM download badge is present but actual download counts are not visible in the provided metadata. Telemetry infrastructure (Sentry) and CI integration suggest real-world usage is anticipated and likely occurring. Adoption not verified at scale from available evidence, though star count and fork count (416) suggest meaningful developer interest beyond casual browsing.
Appears to be a TypeScript CLI tool with a rule-based static analysis engine. Likely uses AST traversal to apply named rules (e.g., 'react-doctor/no-array-index-as-key') across source files. Configuration is handled via a 'doctor.config.ts' file. A GitHub Actions integration suggests a separate CI reporting layer that diffs issues against a baseline. The 'install' command that teaches coding agents likely writes rule descriptions or prompt context into agent configuration files. Architecture appears modular by rule category.
Not documented in README
Last push on 2026-06-23 — the same day as the evaluation date — indicates active, current development. Repository is only ~4 months old and still receiving frequent pushes. Issue tracker is open for contributions. Telemetry via Sentry suggests the team is actively monitoring production usage. Maintenance posture appears healthy and engaged for its age.
ADOPT IF: your team uses AI coding agents to write React and wants a low-friction audit layer to catch agent-generated anti-patterns before they merge; you need a zero-config starting point for React code quality without committing to a full ESLint configuration overhaul. AVOID IF: your codebase already has a comprehensive ESLint+plugin-react+TypeScript setup and disciplined PR review — the marginal value may not justify another tool in the chain; or if you need deep custom rule authoring beyond what the config file supports. MONITOR IF: you are curious about the agent-remediation workflow (the 'install for agents' feature) but need more evidence of rule breadth, false positive rates, and long-term maintenance before committing to CI integration.
Independent dimensions
Mainstream potential
6/10
Technical importance
7/10
Adoption evidence
3/10
- Rule coverage and false-positive rate are not documented in the README — users cannot evaluate analysis quality before trying it, which may cause trust issues in enterprise adoption.
- Created only 4 months ago; longevity and sustained maintenance commitment are unproven, especially if AI-agent coding hype cycles shift.
- The 'agent skill install' feature is novel but unvalidated at scale — it is unclear how reliably it changes agent behavior across different coding agent implementations.
- License is listed as NOASSERTION, meaning the license type could not be auto-detected by GitHub despite the README claiming MIT — this ambiguity may block adoption in regulated or enterprise environments that require license clarity.
- As ESLint plugin-react and TypeScript tooling continue to improve React-specific rules, the differentiation of a separate tool may narrow over time unless React Doctor maintains a meaningful rule quality advantage.
Likely to grow steadily as AI-assisted React development becomes the norm. May settle into a complementary niche alongside ESLint rather than displacing it, particularly if it deepens agent-feedback loop features.
Newsletter
Get analyses like this every Monday
Free weekly digest of the most interesting open-source discoveries.
Languages
Information
- Website
- https://react.doctor
- Language
- TypeScript
- License
- NOASSERTION
- Last updated
- 8h ago
- Created
- 5mo ago
- Analyzed with
- anthropic/claude-haiku-4-5
Stars over time
Contributors over time
Top 100 contributors only — repos with more will plateau at 100.
Open issues
Open pull requests
feat(oxlint-plugin): add `client-prefer-keybind-library` rule
feat(rules): add `no-stale-closure` rule for detecting stale value bugs
feat(oxlint-plugin): add `prefer-schema-validation` draft rule
Top contributors
Similar repos
infinitered/reactotron
Reactotron is a desktop debugging tool for React and React Native applications...
aidenybai/react-scan
React Scan is a developer tool that automatically detects and highlights React...
jsx-eslint/eslint-plugin-react
eslint-plugin-react provides React-specific linting rules for ESLint, helping...
callstackincubator/agent-skills
Agent Skills is a curated collection of AI-optimized prompt instructions and...
| Repository | Stars | Week Δ | Language | Score | Updated |
|---|---|---|---|---|---|
|
millionco/react-doctor
this
|
13.5k | +141 | TypeScript | 7/10 | 8h ago |
|
infinitered/reactotron
|
15.6k | — | TypeScript | 8/10 | 1mo ago |
|
aidenybai/react-scan
|
21.6k | — | TypeScript | 8/10 | 2w ago |
|
jsx-eslint/eslint-plugin-react
|
9.3k | — | JavaScript | 8/10 | 2mo ago |
|
callstackincubator/agent-skills
|
1.5k | — | Shell | 7/10 | 3d ago |
 benjitaylor/agentation
|
4.1k | — | TypeScript | 7/10 | 1mo ago |
The dominant static analysis tool for React. Mature, widely adopted, deeply integrated into editors and CI. React Doctor overlaps in rule coverage but adds a CLI-first, zero-config audit UX and explicitly targets AI-agent remediation workflows. They are more complementary than mutually exclusive — eslint-plugin-react runs continuously in editors, React Doctor is positioned as an audit and agent feedback layer.
react-scan (also from aidenybai/millionco ecosystem) focuses on runtime performance issue detection via visual overlay. React Doctor is static analysis at build/CI time. Different execution model and problem scope — react-scan catches runtime rendering issues, React Doctor catches code pattern issues before runtime.
Reactotron is a runtime debugging and inspection tool for React/React Native apps. It is complementary, not overlapping — it inspects live app state, whereas React Doctor analyzes source code statically.
Biome is a general-purpose linter/formatter for JS/TS with some React rule support. It competes on performance and zero-config setup, but lacks React-specific depth and has no agent-awareness or CI diffing features that React Doctor advertises.
Enterprise-grade code quality platforms with React rule support. Far broader in scope and heavier to integrate. React Doctor is narrower, faster to adopt (npx one-liner), and explicitly React-first. Teams already on Sonar likely see React Doctor as additive rather than a replacement.
