Using GitHub Copilot with GitHub Agentic Workflows

GitHub Agentic Workflows (gh-aw) uses GitHub Copilot as its default AI engine. GitHub Actions runs the Copilot agent from a Markdown workflow, while gh-aw supplies event routing, sandbox controls, and safe outputs for controlled repository writes.

Selection and authentication

Set engine: copilot or omit engine: because Copilot is the default. For organization-billed usage, grant copilot-requests: write; otherwise provide a COPILOT_GITHUB_TOKEN secret containing a fine-grained PAT with Copilot Requests access.

Initialize the repository

Run gh aw init to configure the repository. Copilot is the default engine, so no --engine flag is required. This sets up the full Copilot integration: dispatcher skill, MCP server configuration, .gitattributes, VS Code settings, and the custom agent file.

gh aw init

Example: scheduled repository report

---
on:
  schedule: daily

permissions:
  contents: read
  issues: read
  pull-requests: read
  copilot-requests: write

engine: copilot

safe-outputs:
  create-issue:
    title-prefix: "[status] "
    labels: [report]
    close-older-issues: true
---

# Daily Repository Status

Analyze the repository and create a concise daily status report covering:
- Open issues and their priority
- Recent PR activity
- Upcoming work items

Capabilities and limitations

Copilot supports the broadest set of gh-aw engine-specific features: native custom-agent selection with engine.agent, custom harnesses, max-continuations, bare mode, and per-command bash allowlisting. Copilot CLI does not provide native tools.web-search; configure a supported MCP search integration when the workflow requires web search. See the AI engine feature comparison.

GitHub Agentic Workflows vs. native Copilot assignment

Choose GitHub Agentic Workflows for scheduled or event-driven automation, Markdown-defined workflows, and validated safe outputs. Choose GitHub Copilot’s native @copilot assignment when the task is interactive PR-level coding assistance driven directly from the pull-request conversation.