A Claude Code skill that analyzes any GitHub repository and generates a custom development skill tailored to that project's patterns, conventions, and documentation.
GitHub Cloner examines a repository and extracts:
- Contributor patterns - Who are the key developers? What areas do they focus on?
- Commit conventions - Do they use conventional commits? What prefixes are common?
- PR patterns - How are PRs titled and described? What's typical scope?
- Issue organization - What labels are used? Are there templates?
- Documentation - README, CONTRIBUTING, CLAUDE.md, wiki content
It then generates a skill file that helps any AI assistant work "in tune" with that repository's actual development patterns.
Copy the skill to your Claude Code skills directory:
# Clone this repo
git clone https://github.com/esaruoho/github-cloner.git
# Copy to Claude Code skills directory
cp -r github-cloner ~/.claude/skills/Or manually copy the files to ~/.claude/skills/github-cloner/.
/github-cloner <repository>
Examples:
/github-cloner https://github.com/anthropics/claude-code
/github-cloner facebook/react
/github-cloner owner/repo
- Claude Code CLI
- GitHub CLI (
gh) installed and authenticated
# Install GitHub CLI
brew install gh
# Authenticate
gh auth loginAfter running, a new skill is created at ~/.claude/skills/<repo-name>/ containing:
| File | Purpose |
|---|---|
SKILL.md |
Human-readable skill with conventions, patterns, instructions |
analysis.json |
Machine-readable structured data (for other LLMs) |
- Interactive options - Choose full clone vs API-only, commit history timeframe
- Machine-readable output -
analysis.jsonfollows a JSON schema for LLM interoperability - Cross-LLM compatible - Other AI assistants can consume the structured data
- Incremental updates - Re-run to refresh with latest commits/PRs
| File | Description |
|---|---|
| SKILL.md | The skill instructions (what Claude follows) |
| AGENT-PROTOCOL.md | How other LLMs can consume the output |
| INTEGRATION-GUIDE.md | Building repository-specific skills |
| WORKFLOW-GUIDE.md | Step-by-step usage workflow |
| schemas/repo-analysis-schema.json | JSON schema for analysis.json |
-
Generate a skill from any repo:
cd ~/repos/my-project claude # In Claude: /github-cloner owner/my-project
-
Use the skill from anywhere:
cd ~/dev/my-project-work claude # The skill is automatically available!
-
Update when needed:
/github-cloner owner/my-project
github-cloner/
├── SKILL.md # Main skill instructions
├── README.md # This file
├── AGENT-PROTOCOL.md # LLM interoperability guide
├── INTEGRATION-GUIDE.md # Building on github-cloner
├── WORKFLOW-GUIDE.md # Usage workflow
├── schemas/
│ └── repo-analysis-schema.json
└── scripts/
└── analyze-repo.sh # Helper script
- Onboard to a new codebase - Quickly understand conventions and key contributors
- Cross-LLM workflows - Generate data for Claude, use it with GPT/Gemini
- Build custom skills - Use as a foundation for repository-specific skills
- Team documentation - Auto-generate development guides from actual patterns
MIT
Contributions welcome! Please follow the existing code style and include tests for new features.