ge 0.1.8

Github Expert - Agentic tools for github

ge — GitHub Engineering for AI Agents

Tools and skills that let AI coding agents work on GitHub issues, PRs, and discussions intelligently.

pip install ge
ge install-skills   # register skills with Claude Code

Now, from any project, tell your AI agent "work on THIS" — where THIS can be:

• A GitHub URL: work on https://github.com/owner/repo/issues/42
• A bare number: work on issue #42 (assumes the current repo)
• A repo+number: work on owner/repo#42
• A pre-prepared context folder: work on ~/.cache/ge/owner/repo/issue_42

The agent resolves whatever you give it, confirms what it understood before proceeding, loads or prepares full context (body, comments, media, freshness analysis, AI image descriptions), and works on the issue.

Install skills for Claude Code

ge ships Claude Code skills that teach the agent how to work on GitHub issues. After installing the package, register the skills globally:

ge install-skills

This creates symlinks in ~/.claude/skills/ pointing to the skills bundled with ge. From then on, when you ask Claude Code to work on a GitHub issue in any project, it will automatically:

1. Resolve your input — URL, number, folder path, owner/repo#N — into a target
2. Confirm with you what it resolved before doing any work
3. Load or prepare full context (body, comments, media, cross-references)
4. Analyze freshness — is it stale? already fixed? has related merged PRs?
5. Describe images via the Claude API (or ask you to paste them as fallback)
6. Ask you before working on ambiguous or likely-resolved issues

Three skills are installed:

Skill	Purpose
ge	Full workflow — verify, prepare, analyze, work
ge-analyze	Quick triage — check if an issue is worth working on
ge-context	Context preparation — fetch and assemble structured documents

To remove the skills: ge uninstall-skills

What it does

ge fetches an issue or PR and assembles everything an agent needs:

• Issue/PR body and all comments — with media URLs rewritten to local paths
• Images downloaded — screenshots, mockups, error captures (with auto-detected extensions)
• Video frames extracted — via ffmpeg scene detection; bare GitHub asset URLs are auto-detected as video/image from content
• Freshness analysis — is this issue stale? already fixed? has related merged PRs?
• Cross-references — commits and PRs that mention this issue
• Code file checks — do the files mentioned in the issue still exist?
• AI image descriptions — images are automatically described via the Claude API, so the agent understands visual bugs and UI screenshots without manual pasting

All via gh CLI, so private repos just work.

Quick start

# Install
pip install ge

# Prepare context for an issue
ge prepare owner/repo --number 42

# Or from a URL
ge prepare https://github.com/owner/repo/pull/7

# Just check freshness (no download)
ge analyze-issue owner/repo 42

Requirements

• gh CLI — installed and authenticated (gh auth login)
• ffmpeg — optional, for video frame extraction
• anthropic — optional, for AI-powered image descriptions (pip install anthropic + set ANTHROPIC_API_KEY)
• ImageMagick — optional, for clipboard montage (brew install imagemagick on macOS)
• Python 3.10+

Commands

ge prepare <repo> --number <N>              Full context (auto-detects issue/PR/discussion)
ge prepare <url>                           Full context from a GitHub URL
ge prepare-discussion <repo> --number <N>  Full context for a GitHub Discussion
ge analyze-issue <repo> <N>                Freshness analysis (JSON, no download)
ge analyze-pr <repo> <N>                   PR review analysis with CI status (JSON)
ge fetch-issue <repo> <N>                  Raw issue JSON
ge fetch-pr <repo> <N>                     Raw PR JSON
ge fetch-discussion <repo> <N>             GitHub Discussion JSON
ge media <file.md>                 Download media from markdown
ge video-frames <video>            Extract frames (scene detection by default)
ge describe-images <img>...        Describe images via Claude API (vision)
ge copy-images <img>...            Create montage + copy to clipboard (macOS)
ge resolve <target>                Resolve a URL, folder, or number to structured target
ge install-skills                  Register skills with Claude Code (~/.claude/skills/)
ge uninstall-skills                Remove ge skill symlinks

Project structure

ge/
├── __init__.py      # Facade: prepare(), install_skills(), etc.
├── __main__.py      # CLI via argh (SSOT: _cli_commands list)
├── github.py        # GitHub API via `gh` CLI subprocess
├── media.py         # Image download + video frame extraction (ffmpeg)
├── analysis.py      # Staleness/freshness/relevance analysis
├── context.py       # Assembles everything into context docs
├── util.py          # Internal helpers: gh wrapper, URL parsing, media extraction
└── data/skills/     # Claude Code skills (symlinked by install-skills)
    ├── ge/          # Full workflow skill
    ├── ge-analyze/  # Triage/staleness skill
    └── ge-context/  # Context preparation skill

Python API

import ge

ctx = ge.prepare_issue('owner/repo', 42)
ctx = ge.prepare_pr('owner/repo', 7)
ctx = ge.prepare_discussion('owner/repo', 5)

# Or from any GitHub URL (auto-detects type)
ctx = ge.prepare('https://github.com/owner/repo/issues/42')
ctx = ge.prepare('https://github.com/owner/repo/discussions/5')

# Just analysis
analysis = ge.analyze_issue('owner/repo', 42)

# Resolve flexible input (URL, folder, number, owner/repo#N)
target = ge.resolve_target('#42', current_repo='owner/repo')
# Returns: {repo, number, kind, context_dir, context_md, has_prepared_context, source}

Image analysis

When ge prepare runs, it downloads images and — if anthropic is installed — automatically sends them to the Claude API for visual analysis. The resulting descriptions are embedded in the context document under "Image Descriptions (AI-generated)", so the agent understands screenshots and visual bugs without manual image pasting.

To set up:

pip install anthropic
export ANTHROPIC_API_KEY=sk-ant-...

Image descriptions are generated automatically during ge prepare. To disable them:

ge prepare owner/repo --number 42 -d

Standalone image tools

# Describe images via Claude API
ge describe-images screenshot.png error.jpg

# Describe with a custom prompt
ge describe-images frame1.jpg frame2.jpg --prompt "What changed between these frames?"

# Create montage + copy to clipboard for pasting into Claude Code (macOS, requires ImageMagick)
ge copy-images screenshot1.png screenshot2.png
# Then Cmd+V in Claude Code

Python API:

from ge.media import describe_images, copy_images_to_clipboard

# Get text description of images
text = describe_images('screenshot.png', 'error.jpg')

# Create montage, copy to clipboard
path = copy_images_to_clipboard('img1.png', 'img2.png')