UI はローカライズ済みです。スキルの紹介と説明文書は選択した言語へ AI が自動翻訳し、コード・コマンド・パス・専門用語は原文のままです。原作者の文書が優先されます。
スキルを探す
開発・技術 · SKILL.md

react-doctor

react-doctor

Detects regressions in security, performance, correctness, and architecture across React codebases using score-based quality checks. Surfaces issues with severity ratings and actionable remediation steps for each finding.

Independent GitHub snapshot · Automated classification · Original documentation governs

スキルの説明

GitHub

name: react-doctor description: | Detects regressions in security, performance, correctness, and architecture across React codebases using score-based quality checks. Surfaces issues with severity ratings and actionable remediation steps for each finding. triggers:

  • "react doctor"
  • "React audit"
  • "React performance"
  • "React security"
  • "component quality" das: category: accessibility-quality upstream: "https://github.com/millionco/react-doctor" upstream_path: "SKILL.md" version: latest install: true

react-doctor

Catalogue stub — full skill: millionco/react-doctor

Decision tree

Run this before any React code quality audit:

  1. Is the full skill already installed? Check: ~/.design-agent-skills/skills/react-doctor/SKILL.md exists AND does NOT contain a das: block.

    • Yes → invoke react-doctor and proceed
    • No → go to step 2
  2. Do you have shell access?

    • Yes → run the install command below, then invoke the skill
    • No → show the install command; in Claude Code, send it as a chat message starting with ! — add -g for global install or omit for project-only

Install command

npx skills add millionco/react-doctor

Or per-agent:

  • Claude Code: claude skills add millionco/react-doctor
  • Cursor/OpenCode: npx skills add millionco/react-doctor

Invoke after install

  • Skill name: react-doctor
  • Trigger phrases: "react doctor", "React audit", "React performance", "React security", "component quality"

What it does

Runs score-based quality checks across a React codebase to detect regressions in security, rendering performance, correctness, and architectural soundness. Returns findings ranked by severity with specific component locations and actionable remediation guidance for each issue.

保存したスキル