gh-issues

What This Skill Does

The gh-issues skill is a powerful GitHub issue management tool designed to automate the workflow of identifying, fixing, and integrating changes for GitHub issues. It acts as an orchestrator, leveraging sub-agents to implement fixes for reported bugs or feature requests. The skill fetches issues from a specified repository, optionally filtering them by labels, assignees, or milestones. Once issues are identified, it can automatically spawn sub-agents to develop, test, and commit code changes. It further automates the process of creating pull requests (PRs) from your fork to the main repository and monitors these PRs for review comments. If review comments are detected, it spawns sub-agents to address them, ensuring issues are fully resolved and merged. The skill supports both direct execution and a --watch mode for continuous monitoring, as well as a --dry-run option for testing purposes. It strictly uses the GitHub REST API via curl and does not depend on the gh CLI, making it robust and easily deployable.

Installation

To install the gh-issues skill, run the following command in your OpenClaw environment:

clawhub install openclaw/openclaw/skills/gh-issues

This command will download and configure the skill, making it available for use.

Use Cases

• Automated Bug Fixing: Automatically detect, fix, and create pull requests for issues labeled as 'bug'.
• Continuous Issue Resolution: Monitor a repository for new issues and automatically assign sub-agents to address them.
• PR Review Management: Automatically respond to review comments on your pull requests, accelerating the review process.
• Milestone-based Development: Focus on fixing issues within a specific milestone, such as 'v1.0'.
• Personal Issue Triage: Assign issues to yourself (@me) and have the skill help implement the fixes.
• Forked Repository Workflow: Seamlessly work with your forks, pushing branches and opening PRs from your fork to the upstream repository.
• Scheduled Maintenance: Use --cron mode for automated, unattended runs to catch up on issues or reviews.

Example Prompts

1. gh-issues openclaw/openclaw --label bug --limit 5 --watch --interval 10
2. gh-issues my-org/my-repo --milestone v2.1 --assignee @me --fork my-username/my-repo
3. gh-issues another-org/another-repo --reviews-only --notify-channel -123456789

Tips & Limitations

• GitHub Token: Ensure your GH_TOKEN environment variable is correctly set with appropriate permissions for the gh-issues skill to interact with the GitHub API.
• API Rate Limits: Be mindful of GitHub API rate limits, especially when using --watch mode with a short --interval.
• Fork Setup: For the --fork option to work effectively, ensure your specified fork repository exists and you have push access to it.
• Dry Run: Always use --dry-run first to test your configuration and understand the skill's actions without making any actual changes.
• Error Handling: While the skill is robust, complex issues or unexpected API responses might require manual intervention. Review the output carefully.
• Sub-Agent Configuration: The effectiveness of the fixes depends heavily on the capabilities and configuration of the sub-agents spawned by gh-issues. The --model flag allows specifying a particular LLM for these sub-agents.
• --reviews-only: This flag is extremely useful for focusing solely on addressing feedback for already open PRs without processing new issues.