react-stacklisted
Install: claude install-skill riseshinee/superclaude-react
# react-stack — Stack Detection & Project Conventions
**Step 0** for every `react-*` skill. The goal is to follow the tools and conventions the project actually uses instead of writing code from assumptions.
## 1. Run the detection script
Run `scripts/detect-stack.mjs` from this skill's folder. The path depends on where the skills were installed, so check which one exists:
- Project install: `.claude/skills/react-stack/scripts/detect-stack.mjs`
- Global install: `~/.claude/skills/react-stack/scripts/detect-stack.mjs`
```bash
node .claude/skills/react-stack/scripts/detect-stack.mjs $ARGUMENTS
# add --json for machine-readable output
```
If Node cannot be run, read `package.json`, the lockfile, and config files (`vite.config.*`, `next.config.*`, `tsconfig.json`, `components.json`) directly and fill in the same fields.
## 2. Confirm conventions from real code (what the script can't see)
Dependencies tell you *what* is used, not *how*. Read 2–3 representative files and confirm:
| Item | What to check |
|---|---|
| Component declaration | `function Foo()` vs `const Foo = () =>`, default vs named export |
| Props typing | `interface FooProps` vs `type`, whether `React.FC` is used |
| File naming | `PascalCase.tsx` / `kebab-case.tsx` / `index.tsx` barrels |
| Colocated files | `Foo.test.tsx`, `Foo.stories.tsx`, `Foo.module.css`, `__tests__/` |
| Import paths | path aliases such as `@/`, depth of relative imports |
| Data fetching | custom hooks (`useXxxQuery`), server compo