gh-agent-task

Create GitHub Copilot agent tasks from natural language descriptions via gh agent-task.

4.9k|489|Updated Aug 12, 2025
One-click install
npx skills add https://github.com/githubnext/gh-aw --skill gh-agent-task
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: gh-agent-task
Source: https://github.com/githubnext/gh-aw/tree/main/skills/gh-agent-task
Command: npx skills add https://github.com/githubnext/gh-aw --skill gh-agent-task

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill automates the creation of GitHub Copilot agent tasks—specialized issues that trigger Copilot to perform autonomous code changes—streamlining the delegation of development tasks to AI. It helps AI agents and workflow authors initiate code modifications efficiently.

Core Features & Use Cases

  • Automated Task Creation: Create new agent tasks from natural language descriptions, either inline or from a file, specifying base branches and target repositories.
  • CLI Integration: Leverage gh agent-task commands to create, list, view, and update agent tasks directly from the command line.
  • Safe Output Integration: Seamlessly integrate with gh-aw safe outputs to generate agent tasks from AI agent output.
  • Secure Authentication: Guidance on using Personal Access Tokens (PATs) with appropriate permissions for task creation.
  • Use Case: An AI agent identifies a refactoring opportunity. This skill enables the agent to create a gh agent-task issue with detailed instructions, delegating the code change to GitHub Copilot.

Quick Start

Use the gh-agent-task skill to create a new agent task with the description 'Refactor the user authentication flow in src/auth/login.js'.

Frequently Asked Questions about gh-agent-task

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate code changes using GitHub Copilot agent tasks?

Automate code changes by creating GitHub Copilot agent tasks—specialized issues that trigger Copilot to perform autonomous modifications. Use the `gh agent-task` CLI command with natural language descriptions to delegate code changes across repositories without manual implementation.

Can I create agent tasks from a file instead of typing descriptions inline?

Yes, the `gh agent-task` command supports the `--from-file` flag, allowing you to read task descriptions from a file. This enables integration with AI agent workflows and safe outputs for programmatic task generation.

What permissions do I need to create GitHub Copilot agent tasks?

Creating agent tasks requires elevated permissions: `contents: write`, `issues: write`, and `pull-requests: write`. Authenticate using a Personal Access Token (PAT) or Copilot token with these scopes to generate tasks and produce pull requests.

Does gh agent-task support cross-repository code changes?

Yes, `gh agent-task` handles both single and cross-repository scenarios. Specify the target repository with the `--repo` flag and set the base branch with `--base` to automate code changes across multiple repositories in one workflow.

What output do I get after creating an agent task?

After creation, the command outputs the generated task URL. The agent task issue then triggers Copilot to implement changes and produce a pull request with the automated modifications.

How do I integrate agent tasks with AI agent workflows?

Integrate by combining `gh agent-task` with `gh-aw` safe outputs. This allows AI agents to programmatically generate agent tasks from workflow output, automating the delegation of code changes to GitHub Copilot.