React Doctor Skill: Scan and Fix React Issues With Claude
react-doctor makes Claude scan your React code with the react-doctor tool, keep a 0 to 100 health score from dropping, and fix the issues it reports.
- Installs
- 20.2K
- on skills.sh
- Stars
- 15.1K
- on GitHub
- Files
- 1
- besides SKILL.md
- Updated
- Nov 2026
- last repo commit
After you change React code, the skill has Claude run react-doctor on the changed files and check the score did not drop. If it did, Claude fixes the regressions before committing. For wider cleanup it scans the full project and fixes errors first, then warnings. A design mode runs only UI design checks.
Typing /doctor or asking for a full triage fetches a playbook from react.doctor and follows it. That playbook is remote text that its author can change, so you do not control what it says. SKILL.md says it edits the working tree and never commits or opens pull requests.
It can also record a Chrome performance trace of a running page, with rendering outlined in purple, and explain or tune rules in your doctor.config. Everything runs through npx react-doctor@latest.
Good for
- React, Next.js, Vite and React Native developers who want Claude to check code before it commits
- Teams cleaning up lint, accessibility and bundle-size findings
- Developers tracking down slow renders with a recorded browser trace
Install react-doctor
Claude Code
- 1
Copy this prompt.
Install the react-doctor skill in Claude Code. Run this command: npx react-doctor@latest install Then check that it loaded and tell me how to use it.
- 2
Paste it into Claude Code. It shows the change and asks before running it.
Requires Node.js.
The README says the installer supports Claude Code, Cursor, Codex and OpenCode. Run it from your project root.
From the project README. Source.
What's in the skill
When it loads
Use when finishing a feature, fixing a bug, before committing React code, or when the user types `/doctor`, asks to scan, triage, or clean up React diagnostics. Covers lint, accessibility, bundle size, architecture. Includes a regression check and a full local-triage workflow that fetches the canonical playbook.
Sections in SKILL.md
- After making React code changes:
- For general cleanup or code improvement:
- For a focused UI design audit:
- For runtime performance problems:
- /doctor — full local triage workflow
- Configuring or explaining rules
- Command
- Reuse a base scan
- GitLab warning status
Key facts
- Installs on skills.sh
- 20,172 as of 2026-11-02
- GitHub stars (whole repo)
- 15,072 as of 2026-11-02
- Last repo commit
- 2026-11-01
- Files in the skill folder
- 2 files including SKILL.md
- skills.sh audits
- Gen Agent Trust Hub pass, Socket pass, Snyk warn
What it includes
npx react-doctor@latest --verbose --scope changed
Run after React edits. Reports only issues introduced against the base branch and shows the score.
Runs code ·
Regression checknpx react-doctor@latest --verbose
Scans the full codebase and fixes by severity, errors first.
Runs code ·
Full scannpx react-doctor@latest design --verbose
Runs only design-tagged rules for composition, typography, interaction, accessibility and motion.
Runs code ·
Design auditnpx react-doctor@latest scan <url> --format json
Opens an isolated Chrome profile, records a DevTools trace while you use the page, and stops when you press Enter. Optional --cdp attaches to a running Chrome.
Runs code ·
Runtime traceFetch a playbook from react.doctor and follow it
Uses curl to download the playbook, then runs a scan, filter, triage, fix and validate loop. Rule fix recipes are fetched on demand.
Runs code ·
/doctor triage--scope changed, contextual or lines; --baseline
Controls which findings count. A saved --json scan with --baseline reports only new findings.
Scope optionsrules explain, disable, set, category, ignore-tag
Edits doctor.config.* or package.json#reactDoctor. Described in references/explain.md.
Runs code ·
Rule control
From the official docs
Permissions
High risk
The skill runs npx and, for /doctor, downloads remote instructions and follows them, which can edit your code. Review the diff and approve commands one at a time.
What it can do
- Read your data
- Create and change things
- Delete things
- Read-only mode
Watch out for
/doctor follows a remote playbook
With /doctor, Claude curls https://www.react.doctor/prompts/react-doctor-agent.md and follows every step in it, then fetches per-rule prompts the same way. That text is controlled by the remote site and can change without a reinstall, so Claude may run commands and edit files on instructions you have not read. Read the playbook yourself before the first run and review the diff afterwards. github.com
It runs npx react-doctor@latest every time
The latest version is fetched and run on your machine for each scan. Pin a version if you need repeatable results. github.com
Runtime traces can hold sensitive data
The trace is a local .json.gz file of your app. SKILL.md says not to upload it without permission, and not to copy your browser profile. github.com
The CLI collects crash reports and anonymous usage data
The README says so and gives npx react-doctor@latest --no-telemetry to opt out. github.com
Technical details
| Talks to | npm registry (npx). www.react.doctor (playbook and rule prompts) |
|---|---|
| Scopes |
|
Questions
What does the react-doctor skill do?
It tells Claude to run React Doctor after React changes, keep the 0 to 100 health score from dropping, clean up findings by severity, and run a full /doctor triage when you ask.
How do I install it in Claude Code, Cursor or Codex?
Use the install box above. The README documents npx react-doctor@latest install and says it supports Claude Code, Cursor, Codex and OpenCode. The skills CLI command is listed as an alternative.
Does it need an API key or account?
SKILL.md names none. It runs through npx in your project, and the README mentions a CI install command and documentation site as separate options.
Is it safe to install?
Review what it runs. It executes npx react-doctor@latest and, for /doctor, downloads a remote playbook with curl and follows it, so its instructions can change without a reinstall. The README says the CLI collects crash reports and anonymous usage data, and --no-telemetry opts out.
Does it commit or push code?
The /doctor playbook edits the working tree and, per SKILL.md, never commits or opens pull requests. The regression check tells Claude to fix problems before you commit.
Can it look at runtime performance?
Yes. The scan command with a URL records a Chrome trace while you use the page. SKILL.md says to treat the trace as sensitive local data, never upload it without permission, and never ask for cookies or copy your browser profile.
Information
updated Nov 2, 2026
- Repository
- millionco/react-doctor
- Folder
skills/react-doctor
Sources
Docs, issue trackers and discussions this page is built from.
Show all 5Hide
- react-doctor README · github.com
- skills.sh listing · skills.sh
- react-doctor SKILL.md · github.com
- react-doctor explain.md · github.com
- react-doctor LICENSE · github.com