gh-worker
Automated GitHub issue handling with LLM agents
gh-worker is a CLI tool that automates the entire lifecycle of GitHub issue resolution, from syncing issues to generating implementation plans and creating pull requests using LLM agents.
Overview
gh-worker streamlines software development workflows by automating routine issue handling tasks:
- Sync issues from GitHub to local storage
- Plan implementations using AI agents (Claude Code, OpenCode, Gemini, Codex)
- Implement plans automatically and create pull requests
- Monitor ongoing implementations in real-time
Key Features
Automated Issue Processing
Automatically sync GitHub issues to local files, making them available for processing by LLM agents.
AI-Powered Planning
Generate comprehensive implementation plans using state-of-the-art LLM agents that understand your codebase and issue requirements.
Parallel Execution
Process multiple issues concurrently with configurable parallelism for both planning and implementation phases.
Continuous Workflow
Run sync → plan → implement cycles continuously with configurable frequencies, or execute on-demand for specific issues.
Pluggable Architecture
Support for multiple LLM agents through a flexible plugin system. Ships with built-in support for:
- Claude Code (default)
- OpenCode
- Google Gemini
- OpenAI Codex
File-Based Storage
Simple, transparent file-based storage system that makes it easy to inspect, version control, and manage issue data and plans.
Robust Error Handling
Automatic retry logic with exponential backoff for transient failures, ensuring reliable operation in production environments.
Quick Start
# Install with uv
uv tool install https://github.com/tomzxcode/gh-worker.git
# Configure
ghw config issues-path ~/gh-worker/issues
ghw config repository-path ~/gh-worker/repos
# Add a repository
ghw add owner/repo
# Sync issues
ghw sync --repo owner/repo
# Generate plans
ghw plan --repo owner/repo
# Implement
ghw implement --repo owner/repo
For detailed instructions, see the Installation and Usage guides.
Architecture
gh-worker is built with a modular architecture:
- CLI Layer: Built with cyclopts for a clean command-line interface
- Storage Layer: File-based storage for issues and plans
- Agent Layer: Pluggable LLM agent system
- Execution Layer: Parallel execution orchestration
- GitHub Layer: Wrapper around GitHub CLI for repository operations
For technical details, see the Architecture documentation.
Use Cases
Automated Issue Triage
Automatically generate implementation plans for new issues, helping maintainers quickly understand the scope and complexity of bug reports and feature requests.
Batch Processing
Process multiple issues in parallel, ideal for cleaning up issue backlogs or handling routine maintenance tasks.
Continuous Integration
Run gh-worker continuously to automatically handle issues as they arrive, creating a fully automated issue-to-PR pipeline.
Development Assistance
Use gh-worker as a development assistant that generates implementation plans you can review and modify before execution.
Next Steps
- Installation Guide - Get gh-worker up and running
- Configuration - Configure gh-worker for your workflow
- Usage Guide - Learn the commands and workflows
- Agents - Understand the agent system
- Architecture - Technical deep dive
Requirements
- Python 3.12 or later
- uv package manager
- GitHub CLI authenticated with your account
- Claude Code CLI (for the default agent)
License
MIT License - See LICENSE for details.
Support
For issues, questions, or contributions:
- GitHub Issues: gh-worker issues
- Repository: gh-worker