| GitHub repository code analysis. Uses local script `skills/scripts/github_analyze.js` to convert a repo into an AI-friendly Markdown snapshot. No external API required.
Install with the open skills CLI (global, non-interactive — available in every Claude Code session):
npx skills add LeoYeAI/openclaw-master-skills --skill "github-analysis" -g -a claude-code -yOr manually — copy the SKILL.md below into:
~/.claude/skills/github-analysis/SKILL.mdPart of the Crab Catch skill collection — installing the parent includes this skill.
---
name: github-analysis
description: |
GitHub repository code analysis. Uses local script
`skills/scripts/github_analyze.js` to convert a repo into an AI-friendly
Markdown snapshot. No external API required.
metadata:
author: NotevenDe
version: 1.2.0
---
# GitHub Analysis — Repository Inspection
## Local Script
Script path: `skills/scripts/github_analyze.js`
### Functions
| Function | Description |
|----------|-------------|
| `analyzeRepository(url, options?)` | Metadata, structure, README, languages, architecture, recent commits |
| `convertToMarkdown(url, options?)` | Convert repo to a single AI-friendly Markdown document with file contents |
| `parseGitHubUrl(url)` | Parse GitHub URL or `owner/repo` into `{ owner, repo }` |
### Options
| Option | Default | Description |
|--------|---------|-------------|
| `depth` | 2 | Directory depth limit (analyzeRepository only) |
| `maxFiles` | 75 | Max files to include (convertToMarkdown only) |
| `maxSize` | 30000 | Max size per file in bytes (convertToMarkdown only) |
## Fallback Behavior
| Scenario | Behavior |
|-----------------------------|-------------------------------------------|
| Repo not found or private | Skip, record a warning |
Use when completing tasks, implementing major features, or before merging to verify work meets requirements
Use when implementing any feature or bugfix, before writing implementation code
Use when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always