Repolode
Submit
github

github/gh-aw

GitHub CLI extension that lets you write AI-powered repository automation as Markdown and runs it securely through GitHub Actions.

View on GitHubWebsite
Stars
5,422
Forks
596
Open issues
339
Last commit
5 hours ago

What is it?

GitHub Agentic Workflows (gh-aw) lets developers define AI agent tasks for a repository in Markdown files with YAML frontmatter. The gh aw compile command turns each workflow into a standard GitHub Actions workflow file that runs on GitHub.

It is meant for tasks that need reasoning or interpretation, such as issue triage, pull request review, CI failure investigation, documentation upkeep, dependency analysis, and repository reporting. Deterministic builds, tests and deployments stay in ordinary Actions workflows; agentic workflows sit alongside them.

Built-in engines include GitHub Copilot, Claude Code, OpenAI Codex, Google Gemini and Pi. Agent jobs are read-only and sandboxed by default, and GitHub writes are applied through validated safe-outputs jobs with scoped permissions.

Key features

  • 01Defines workflows in Markdown with YAML frontmatter for triggers, permissions, tools and engine choice.
  • 02Compiles each workflow into a standard GitHub Actions workflow with gh aw compile.
  • 03Supports GitHub Copilot, Claude Code, OpenAI Codex, Google Gemini and Pi engines.
  • 04Runs agent jobs read-only and sandboxed by default.
  • 05Applies GitHub writes through validated safe-outputs jobs with scoped permissions.

Use cases

  • Triaging new issues and labeling them automatically.
  • Reviewing pull requests with an AI agent as part of CI.
  • Investigating failed CI runs and summarizing the cause.
  • Keeping documentation up to date and generating periodic repository reports.

Quick start

  1. 1

    Install the GitHub CLI extension

    gh extension install github/gh-aw
  2. 2

    Follow the quickstart to select an AI engine, add a sample workflow, and run it through GitHub Actions

Things to know

  • Needs an AI engine account or API key, such as GitHub Copilot, Claude Code, OpenAI Codex, or Google Gemini.
  • Agentic workflows require careful human supervision, and the README says to use it at your own risk.
  • Versions 0.83.3 up to but not including 0.85.4 were retired after a security vulnerability was found.

Topics

#github-actions#ai-agents#automation#ci#llm#github-cli#copilot#claude#actions#cai

FAQ

What is gh-aw?
GitHub CLI extension that lets you write AI-powered repository automation as Markdown and runs it securely through GitHub Actions. GitHub Agentic Workflows (gh-aw) lets developers define AI agent tasks for a repository in Markdown files with YAML frontmatter. The gh aw compile command turns each workflow into a standard GitHub Actions workflow file that runs on GitHub.
Is gh-aw free to use? What license does it have?
gh-aw is open source under the MIT license. Check the license terms before using it in commercial products.
Is gh-aw still maintained?
Actively maintained. The last commit to gh-aw was on Oct 20, 2026.
What can I use gh-aw for?
Triaging new issues and labeling them automatically. Reviewing pull requests with an AI agent as part of CI. Investigating failed CI runs and summarizing the cause. Keeping documentation up to date and generating periodic repository reports.
How do I install gh-aw?
Install the GitHub CLI extension: gh extension install github/gh-aw Follow the quickstart to select an AI engine, add a sample workflow, and run it through GitHub Actions
Where can I get gh-aw?
The source code and installation instructions are on GitHub: https://github.com/github/gh-aw

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects