create-agent-skills

Create Claude Code skills and slash commands with SKILL.md structure.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/mattjefferson/open-compound-engineering --skill create-agent-skills-mattjefferson
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: create-agent-skills
Source: https://github.com/mattjefferson/open-compound-engineering/tree/main/skills/create-agent-skills
Command: npx skills add https://github.com/mattjefferson/open-compound-engineering --skill create-agent-skills-mattjefferson

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This guide helps skill authors create Claude Code skills and slash commands that conform to official specifications, enabling consistent tooling and reusable patterns.

Core Features & Use Cases

  • Frontmatter-driven metadata: standardize name, description, and invocation controls
  • Router vs single-file skills: decide when to split into workflows, references, and scripts
  • Progressive disclosure: keep SKILL.md concise and load knowledge on demand
  • Examples and templates: provide starter patterns for SKILL.md, workflows, and references
  • Essential principles: inline requirements to ensure safety and consistency across models
  • Slash-command integration: align with Claude Code commands by merging commands and skills when needed

Quick Start

Create a new skill directory with SKILL.md following the examples, then add at least one workflow and a references file to demonstrate progressive disclosure.

Frequently Asked Questions about create-agent-skills

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

FAQPage Schema
How do I create Claude Code skills with standard frontmatter and routing?▼

To create Claude Code skills, you compose a SKILL.md file with standard frontmatter for metadata, then implement XML-based routing to split workflows, references, and scripts for scalable and consistent tooling.

What is progressive disclosure in skill authoring and when should I use it?▼

Progressive disclosure in skill authoring keeps SKILL.md concise by loading knowledge on demand. Use it to ensure scalable skills by separating core routing logic from detailed workflow references and scripts.

How do I structure a SKILL.md file to align with Claude Code slash commands?▼

Structure SKILL.md using standard frontmatter, XML-based structure, and intake routing. Align with Claude Code slash commands by merging command definitions and skill workflows to ensure consistent invocation controls.

Does this skill authoring guide support single-file skills or do I need multiple workflow files?▼

This skill authoring guide supports both router and single-file skills. You decide when to split into workflows, references, and scripts based on complexity, ensuring progressive disclosure and reusable patterns.

What's the best way to manage references when authoring new skills?▼

The best way to manage references when authoring skills is to enforce standard reference management by separating detailed knowledge into dedicated reference files, keeping the main SKILL.md concise and loading context on demand.

Can I refine existing skills without breaking their frontmatter and invocation controls?▼

Yes, you can refine existing skills by updating workflows and references while enforcing standard frontmatter and invocation controls, ensuring the updated skill structure maintains essential principles and consistency across models.