setup
Skill2 starsChanged 9 months agoMITView on GitHub
Analyze the current project and install claude-react-kit customized for its actual stack, structure, and conventions
191 other projects use a version of this.
Adopted by
Most recently taken up by
- hydra-db/hydradb-claude-codetheir own version, first seen Nov 8, 2026
- WALKERKILLER/Look-Tongji-Notestheir own version, first seen Nov 8, 2026
- steerlabs/opensteertheir own version, first seen Nov 8, 2026
- ansvisor/ansvisortheir own version, first seen Nov 8, 2026
- dsifry/metaswarmtheir own version, first seen Nov 8, 2026
- anombyte93/prd-taskmastertheir own version, first seen Nov 8, 2026
Used with
Projects with it also use add-voice-transcription, add-discord and 10 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
juancruzmunozalbeloSkilladd-voice-transcription
Add voice message transcription to NanoClaw using OpenAI's Whisper API. Automatically transcribes WhatsApp voice notes so the agent can read and respond to them.
Also in 5 projects that use this
MicrosoftSkilladd-discord
Add Discord bot channel integration via Chat SDK.
Also in 11 projects that use this
MicrosoftSkilladd-slack
Add Slack channel integration via Chat SDK.
Also in 11 projects that use this
juancruzmunozalbeloSkilladd-telegram-swarm
Add Agent Swarm (Teams) support to Telegram. Each subagent gets its own bot identity in the group. Requires Telegram channel to be set up first (use /add-telegram). Triggers on "agent swarm", "agent teams telegram", "telegram swarm", "bot pool".
Also in 7 projects that use this
- +8 more from microsoft
Other versions191 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 191
- anthropics/knowledge-work-plugins★ 35ksame name
- garrytan/gbrain★ 32ksame name
- gavrielc/nanoclaw★ 31ksame name
- nanocoai/nanoclaw★ 31ksame name
- qwibitai/nanoclaw★ 31ksame name
- Q00/ouroboros★ 6.3ksame name
- pytorch/executorch★ 5.2ksame name
- agenticnotetaking/arscontexta★ 3.5ksame name
- aws/agent-toolkit-for-aws★ 3.4ksame name
- spotify/portal-ai-plugins★ 2.8ksame name
181 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository3 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- audit-architectureSkill
- scaffold-featureSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
35 rules. 0 are common, 25 are distinctive, and 10 are very specific to this project.
0
Common
Most of the 171 edited versions we read keep these.
25
Distinctive
This author's own choices. Worth a look.
10
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 33
After analysis, present your findings to the user in a concise summary:
- Distinctive: line 48
If the architecture doesn't clearly match the 5-layer model, ask how they want to map their directories
- Distinctive: line 49
If there are conventions you noticed but aren't sure if they're intentional
- Distinctive: line 50
If there's an existing CLAUDE.md, ask what to keep vs replace
- Distinctive: line 53
Do NOT ask about things you already know from reading the code.
- Distinctive: line 64
If they had an existing CLAUDE.md, merge in any project-specific rules that don't conflict
- Distinctive: line 71
All hooks wired up with correct paths
- Distinctive: line 76
Use their real directory names (not generic ones)
- Very specific: line 77
If they use app/ (Next.js/Expo), reference that as the route layer
- Very specific: line 78
If they use pages/, screens/, or something else, adapt accordingly
- Very specific: line 79
If they have a lib/ or utils/ directory, map it into the layer model
- Very specific: line 85
If they use CSS Modules → reference .module.css files in the component structure
- Distinctive: line 86
If they use Tailwind → no separate styles file, mention className patterns
- Very specific: line 87
If they use styled-components → reference styles.ts with styled exports
- Very specific: line 88
If they use React Native StyleSheet → reference styles.ts with StyleSheet.create
- Distinctive: line 89
If they use Storybook → mention story files in component structure
- Distinctive: line 95
If React Query → add query key conventions, hook patterns from their code
- Distinctive: line 96
If Zustand → add store patterns, useShallow if they use it
- Distinctive: line 97
If Redux/RTK → add slice conventions, selector patterns
- Distinctive: line 98
If Apollo → add fragment/query patterns
- Distinctive: line 99
If they use multiple → document how each is used (server state vs client state)
- Distinctive: line 103
Keep this mostly generic but adapt:
- Distinctive: line 104
If they use barrel files extensively, don't ban them — warn about deep chains
- Distinctive: line 105
If they have a monorepo, add cross-package import rules
- Very specific: line 112
Check if their hooks follow use prefix consistently
- Distinctive: line 113
Check their service/API file naming
- Distinctive: line 114
Document what they ACTUALLY do, not an ideal — unless they ask for a migration
- Distinctive: line 118
Keep this generic — it's about Claude Code behavior, not project-specific.
- Distinctive: line 136
After generating all files, show the user a summary of what was created:
- Very specific: line 143
NEVER use placeholder values like YOUR_COMMAND_HERE — always use real detected values
- Distinctive: line 144
NEVER contradict their existing ESLint/Prettier configuration
- Distinctive: line 145
NEVER assume a directory exists — check first
- Distinctive: line 146
If the project doesn't have a clear layered architecture yet, set up the rules as aspirational but note this in the CLAUDE.md
- Very specific: line 148
If .claude/ already exists, MERGE intelligently — don't overwrite custom rules the user may have added
- Very specific: line 149
Keep CLAUDE.md under 70 lines — move details to .claude/rules/
“Common” is measured: kept by at least 6 in 10 of the 171 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
claude-react-kit — Setup
You are setting up claude-react-kit in this project. Your job is to analyze the codebase, understand its real stack and conventions, and generate perfectly customized configuration files.
Phase 1: Deep Analysis
Analyze the project thoroughly. Read and extract information from:
package.json— framework, dependencies, scripts, package manager (check forpackageManagerfield too)- Lock file — which exists? (
package-lock.json,yarn.lock,pnpm-lock.yaml,bun.lock,bun.lockb) tsconfig.json— path aliases, strict mode, baseUrl, target- Framework config —
next.config.*,vite.config.*,app.json/app.config.*(Expo),metro.config.*,remix.config.* - Linter/formatter —
.eslintrc*,eslint.config.*,.prettierrc*,biome.json - Existing
CLAUDE.md— if one exists, read it to preserve any project-specific rules the user already has - Existing
.claude/— if it exists, read settings.json and any custom rules to preserve them - Source directory structure —
ls src/or equivalent to understand the current architecture (do they already have features/, shared/, core/, design-system/, or different names?) - Routing —
ls app/orls pages/orls src/screens/to understand routing structure - State management patterns — grep for store/slice/atom definitions to understand actual usage
- Test setup —
jest.config.*,vitest.config.*, test file patterns - CI/CD —
.github/workflows/, check what checks run
Build a complete mental model of the project before writing anything.
Phase 2: Ask What You Can't Detect
After analysis, present your findings to the user in a concise summary:
Here's what I detected:
- Framework: [X]
- Package manager: [X]
- Source structure: [describe what you found]
- State management: [X]
- Testing: [X]
- Styling: [X]
- Existing CLAUDE.md: [yes/no, summary if yes]
- Path aliases: [X]Then ask ONLY about things you genuinely couldn't determine:
- If the architecture doesn't clearly match the 5-layer model, ask how they want to map their directories
- If there are conventions you noticed but aren't sure if they're intentional
- If there's an existing CLAUDE.md, ask what to keep vs replace
- Any ambiguities in their setup
Do NOT ask about things you already know from reading the code. Be efficient.
Phase 3: Generate Files
3.1: CLAUDE.md
Generate a CLAUDE.md that is:
- Under 70 lines
- Has the REAL commands from their
package.jsonscripts (not placeholders) - Architecture diagram adapted to their ACTUAL directory structure and framework
- References their real path aliases
- If they had an existing CLAUDE.md, merge in any project-specific rules that don't conflict
3.2: .claude/settings.json
Generate settings with:
- Permission
allowlist using their ACTUAL package manager and script names - Permission
denylist for their sensitive files - All hooks wired up with correct paths
3.3: .claude/rules/architecture.md
Customize based on their ACTUAL structure:
- Use their real directory names (not generic ones)
- If they use
app/(Next.js/Expo), reference that as the route layer - If they use
pages/,screens/, or something else, adapt accordingly - If they have a
lib/orutils/directory, map it into the layer model - Define boundaries based on what directories ACTUALLY exist
3.4: .claude/rules/components.md
Adapt to their conventions:
- If they use CSS Modules → reference
.module.cssfiles in the component structure - If they use Tailwind → no separate styles file, mention className patterns
- If they use styled-components → reference
styles.tswith styled exports - If they use React Native StyleSheet → reference
styles.tswith StyleSheet.create - If they use Storybook → mention story files in component structure
- Detect their actual component size patterns and set realistic limits
3.5: .claude/rules/state-management.md
Customize based on what they ACTUALLY use:
- If React Query → add query key conventions, hook patterns from their code
- If Zustand → add store patterns, useShallow if they use it
- If Redux/RTK → add slice conventions, selector patterns
- If Apollo → add fragment/query patterns
- If they use multiple → document how each is used (server state vs client state)
3.6: .claude/rules/code-quality.md
Keep this mostly generic but adapt:
- If they use barrel files extensively, don't ban them — warn about deep chains
- If they have a monorepo, add cross-package import rules
- Match their existing lint rules (don't contradict ESLint config)
3.7: .claude/rules/naming-conventions.md
Extract from their ACTUAL code:
- Scan a few directories to detect whether they use PascalCase, kebab-case, or camelCase for component folders
- Check if their hooks follow
useprefix consistently - Check their service/API file naming
- Document what they ACTUALLY do, not an ideal — unless they ask for a migration
3.8: .claude/rules/session-management.md
Keep this generic — it's about Claude Code behavior, not project-specific.
3.9: .claude/hooks/
Generate all 4 hooks:
validate-architecture.sh— use their REAL directory names in the boundary checksprotect-files.sh— use their REAL lock file and framework-specific protected pathspost-compact-context.sh— summarize THEIR architecture rules, not generic onespre-compact-save.sh— this one is generic, keep as-is
3.10: .claude/skills/
Install both skills:
audit-architecture/SKILL.md— adapt grep patterns to their real directory structurescaffold-feature/SKILL.md— adapt the scaffold to their conventions (CSS approach, test files, etc.)
Phase 4: Present and Confirm
After generating all files, show the user a summary of what was created:
- List every file with a one-line description of what was customized
- Highlight any decisions you made that they might want to review
- Suggest they run
/audit-architectureto see their current state
Important Rules
- NEVER use placeholder values like
YOUR_COMMAND_HERE— always use real detected values - NEVER contradict their existing ESLint/Prettier configuration
- NEVER assume a directory exists — check first
- If the project doesn't have a clear layered architecture yet, set up the rules as aspirational but note this in the CLAUDE.md
- Preserve any existing
.claude/settings.local.json— never touch it - If
.claude/already exists, MERGE intelligently — don't overwrite custom rules the user may have added - Keep CLAUDE.md under 70 lines — move details to .claude/rules/
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

