An advanced orchestrator skill that fetches GitHub issues, spawns parallel AI sub-agents to implement fixes, and automatically addresses PR review comments.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install cool-gh-issues
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install cool-gh-issues using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is gh-issues?
The gh-issues skill is a high-level orchestration tool designed for Openclaw Skills users who want to automate the entire lifecycle of software maintenance. It acts as a project manager, identifying actionable bugs or enhancements from a GitHub repository and delegating the actual coding work to specialized, parallel sub-agents. By leveraging the GitHub REST API and local git operations, it ensures that fixes are implemented in isolated branches without manual intervention.
Beyond just opening pull requests, this skill features a sophisticated PR Review Handler. It monitors existing pull requests for human feedback or AI-generated reviews, analyzes comments for actionability, and re-spawns agents to address requested changes. This creates a continuous loop of improvement, significantly reducing the manual burden on maintainers and accelerating development velocity.
gh-issues Use Cases
- Automatically processing a backlog of bug reports by spawning parallel agents to find and fix root causes.
- Running in a cron-based environment to periodically check for new issues and initiate fixes overnight.
- Monitoring open pull requests to automatically address minor review comments or formatting requests.
- Scaling open-source maintenance by using fork mode to contribute fixes from a personal repository to an upstream source.
How gh-issues Works
- The skill parses user arguments to define the target repository, labels, milestones, and operational modes like fork or dry-run.
- It fetches eligible issues via the GitHub REST API, filtering out pull requests and already processed items.
- Pre-flight checks verify the local git environment, remote access, and ensure no duplicate PRs or in-progress branches exist for the selected issues.
- It spawns parallel sub-agents (up to 8 concurrently), providing each with a specific task prompt to analyze code, implement a fix, and run tests.
- Sub-agents push their changes to the designated remote and create a detailed Pull Request with a summary of changes and test results.
- The orchestrator enters a review handling phase, checking for new comments on open PRs and spawning sub-agents to resolve reviewer feedback.
gh-issues Setup
To use this skill, ensure you have a GitHub personal access token (GH_TOKEN) configured in your environment or Openclaw config. The skill requires curl, git, and the gh CLI for initial authentication.
# Standard usage to process bugs in the current repo
/gh-issues --label bug --limit 5
# Advanced usage with a fork and a specific AI model
/gh-issues owner/repo --fork user/repo --model glm-5 --watch --interval 10
gh-issues Data Schema & Taxonomy
The skill maintains state across operations using a structured file system in the .clawdbot directory:
| File/Component | Description |
|---|---|
gh-issues-claims.json |
Tracks active sub-agent assignments with timestamps to prevent duplicate work. |
gh-issues-cursor-{repo}.json |
Stores the last processed issue ID for sequential cron-mode operations. |
PROCESSED_ISSUES |
An in-memory set (persisted during watch mode) to avoid re-fetching handled issues. |
ADDRESSED_COMMENTS |
A registry of comment IDs that have already been responded to by the AI. |
gh-issues Advanced Features
- Parallel Execution: Supports up to 8 concurrent sub-agents to maximize throughput.
- Fork Mode: Safely develops fixes in a personal fork while targeting the upstream repository for PRs.
- Watch & Poll: Continuously monitors repositories for new issues and review comments at specified intervals.
- Telegram Notifications: Sends real-time status updates and PR links to a designated Telegram channel.
- Confidence Scoring: Sub-agents self-assess issue actionability and skip tasks with vague requirements to save tokens.
SKILL.md
Loading
Related Openclaw Skills
AI Instagram Bio Generator
An intelligent automation tool that generates 10 conversion-optimized Instagram bios tailored to your niche using competitor analysis and AI scoring.
AI Podcast Show Notes Generator
An automated workflow that generates ten SEO-optimized show note versions, timestamps, and social media promotion kits for podcast episodes.
Editor Free Download
An AI-powered video editing skill that transforms raw recordings into polished 1080p MP4 files through simple text commands without requiring local software.
AI Newsletter Subject Line Generator
An AI-powered optimization tool that generates 15 high-conversion newsletter subject lines using proven formulas and predictive open-rate scoring.
Corkboard Dashboard
A real-time visual dashboard for managing actionable pins, multi-track project pipelines, and historical task recovery.
Node Connect
A comprehensive diagnostic tool for resolving connection, pairing, and routing failures between OpenClaw gateways and companion apps.
Featured*
