name	description	metadata
github-triage	Daily check any GitHub repo for new issues, analyze with AI reasoning, and create detailed implementation reports in Obsidian. Works with any repository.	clawdbot emoji requires 🔍 bins env gh GITHUB_TOKEN

github-triage: AI-Powered GitHub Issue Triager

Daily automation to check, analyze, and document issues in any GitHub repository using AI reasoning.

How It Works

This skill uses Claude's reasoning to:

1. Clone/pull the repo to ~/workspace/{repo-name}
2. Check for new issues since last run (using tracking file)
3. Read and understand each issue in depth
4. Investigate the codebase to understand implementation requirements
5. Create detailed analysis reports with:
   • Issue summary and context
   • Implementation feasibility assessment
   • Suggested code changes with examples
   • Configuration examples (if applicable)
   • Testing recommendations
   • Next steps for implementation
6. Track processed issues to avoid duplicates

Usage

Tell me: "Run the github-triage skill for owner/repo"

Examples:

• "Run github-triage for rust-lang/rust"
• "Check issues for vercel/next.js"
• "Triage issues on my-org/my-project"

Output

• Analysis reports: ~/Documents/personal/inbox/{repo}-issue-{number}.md
• Tracking file: ~/.clawdbot/skills/github-triage/.last-check.json
• Repository checkout: ~/workspace/{repo-name}/

Issue Classification (AI-Driven)

I analyze each issue using judgment:

• IMPLEMENTABLE: Clear bug or small enhancement
• NEEDSDiscussion: Requires clarification or architectural decision
• OutOfScope: Feature request too large/complex
• NotFixable: Cannot reproduce or insufficient info

Before Investigating

• Run git pull to get latest changes
• Use the branch mentioned in the issue, or the repo's default branch

Report Contents

Each report includes:

1. Issue metadata (author, date, labels, type)
2. Summary of the issue
3. Analysis of current codebase
4. Implementation assessment (fixable, complexity, approach)
5. Suggested file changes with code examples
6. Configuration examples (if applicable)
7. Testing recommendations
8. Next steps checklist

Setup

Ensure GitHub CLI is authenticated:

gh auth login

Or set GITHUB_TOKEN environment variable.

Tracking

The skill tracks:

• Timestamp of last check
• List of processed issue numbers per repository

New issues (not in processed list) are triaged. Old issues are skipped.

Notes

• Each repository has its own tracking file entry
• Reports are named after the repository to avoid conflicts
• Works with any public or private GitHub repository you have access to