custom-agents

Configure GitHub Copilot custom agent files with YAML frontmatter schemas.

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

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill empowers users to define specialized AI agent behaviors and tool access through custom agent files, ensuring AI agents perform tasks precisely according to repository-specific standards and requirements. It helps optimize AI performance for specific development workflows.

Core Features & Use Cases

  • Flexible Agent Scoping: Define instructions repository-wide, path-specific, or as custom agent profiles for task-specific workflows.
  • Granular Tool Control: Specify exactly which tools (e.g., read, edit, search, pr) an agent is allowed to use, enhancing security and focus.
  • Custom Instructions: Provide natural language prompts and guidelines to shape the agent's behavior and output style.
  • Integration with gh-aw: Seamlessly integrate custom agents into GitHub Agentic Workflows via the imports field.
  • Use Case: A team wants an AI agent to strictly follow frontend development guidelines. This skill guides them to create a .github/instructions/frontend.instructions.md file with applyTo: "src/frontend/**" and specific instructions.

Quick Start

Consult the custom-agents skill to create a readme-creator.md agent profile that specializes in generating clear and comprehensive README files.

Frequently Asked Questions about custom-agents

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

FAQPage Schema
How do I configure AI agents to follow repository-specific guidelines?

Configure repository-specific AI agents using custom agent files in YAML frontmatter format with name, instructions, and tool specifications. Define agents repository-wide, path-specific, or as custom profiles to enforce your development standards and control which tools agents can access.

What tools can I specify for custom GitHub agents?

Custom agents support granular tool control including read, edit, search, and pr tools. Specify exactly which tools an agent is allowed to use through the tools field in your agent configuration to enhance security and focus agent behavior on required tasks.

Can I apply different agent instructions to specific code paths?

Yes, use path-specific instructions with the applyTo field to target agent behavior to particular directories or file patterns. This enables specialized agents—like frontend-specific or documentation-focused agents—to apply tailored rules only where needed.

How do custom agents integrate with GitHub Copilot workflows?

Custom agents integrate with GitHub Agentic Workflows through the imports field in your agent configuration. This seamless integration lets you deploy tailored AI agent behaviors directly into your development workflows without additional setup.

What's the file structure and schema for custom agent configurations?

Custom agents use a single YAML frontmatter file per workflow with required name field and optional description, prompt, tools, applyTo, MCP server, and settings fields. Store these in .github/instructions/ to apply repository-scoped or workflow-scoped behavior.

Can I create task-specific agent profiles for specialized workflows?

Yes, define custom agent profiles for task-specific workflows like README generation, test suite creation, code reviews, and documentation improvements. Each profile combines targeted instructions and tool access to optimize AI performance for distinct development tasks.