RESEARCH PROTOTYPE
GitHub Agentic Workflows

Automate GitHub tasks safely using AI agents—write workflows in plain English, no YAML required.
AI agents execute complex GitHub operations like issue triage, PR reviews, and repository maintenance automatically.

Getting Started

What are GitHub Agentic Workflows?

GitHub Agentic Workflows allow you to write repository-level AI agent descriptions in plain markdown and run automatically in GitHub Actions. Describe your repository automation goals and execute safely with built-in security controls, having AI make intelligent decisions based on context.

How it works:

Write - Create a .md file with your automation instructions in natural language
Compile - Run gh aw compile to transform it into a secure GitHub Actions workflow (.lock.yml)
Run - GitHub Actions executes your workflow automatically based on your triggers

Security Built-In

Workflows run with read-only permissions by default. Write operations require explicit approval through sanitized safe outputs (pre-approved GitHub operations), with sandboxed execution, tool allowlisting, and network isolation ensuring AI agents operate safely within controlled boundaries.

Key Features

Natural Language Workflows

Write automation in markdown instead of complex YAML

AI-Powered Decision Making

Workflows that understand context and adapt to situations

GitHub Integration

Deep integration with Actions, issues, PRs, discussions, and repository management

Safety First

Sandboxed execution with minimal permissions and safe output processing

Multiple AI Engines

Support for Claude, Codex, and custom AI processors

Continuous AI

Systematic, automated application of AI to software collaboration

Example: Daily Issues Report

Here’s a simple workflow that runs daily to analyze repository issues:

---
on: daily
permissions: read
safe-outputs:
  create-discussion:
---

## Daily Issues Report

Analyze repository issues and create a discussion
with metrics, trends, and key insights.