create-agent-skills

Guide skill authors to design, document, and validate Claude Code skills and slash commands.

10|Updated Feb 15, 2026
One-click install
npx skills add https://github.com/aimi-so/aimi-engineering-plugin --skill create-agent-skills-aimi-so
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: create-agent-skills
Source: https://github.com/aimi-so/aimi-engineering-plugin/tree/main/plugins/aimi-engineering/skills/create-agent-skills
Command: npx skills add https://github.com/aimi-so/aimi-engineering-plugin --skill create-agent-skills-aimi-so

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Guides skill authors to design, document, and validate Claude Code skills and slash commands to ensure consistent structure and reliable behavior.

Core Features & Use Cases

  • Clear frontmatter guidance (name, description, optional fields) and consistent naming
  • Router-pattern structure for multi-workflow skills with intake and routing
  • Progressive disclosure: workflows, references, and scripts to keep SKILL.md concise
  • Validation, best-practices, anti-patterns, and naming conventions
  • Reference/indexing: reference_index and workflows_index for discoverability

Quick Start

Follow the template in this guide to create a new skill using the router pattern.

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 consistent structure and reliable behavior?

To create Claude Code skills, follow a structured template that enforces clear frontmatter, consistent naming conventions, and validation rules. This ensures reliable execution and consistent discovery for your slash commands.

What is the router pattern for designing multi-workflow Claude Code skills?

The router pattern structures multi-workflow skills by directing an initial intake process to specific workflows. It uses progressive disclosure with separate references and scripts to keep the main SKILL.md file concise.

How do I use progressive disclosure when authoring Claude Code slash commands?

Progressive disclosure separates workflows, references, and scripts into indexed files rather than embedding everything in SKILL.md. This keeps the core skill concise while maintaining deep functionality through reference_index and workflows_index.

What frontmatter fields are required when building a Claude Code skill?

Required frontmatter fields for a Claude Code skill include the name and description, while optional fields provide additional metadata. Enforcing these fields ensures high-quality skills with proper validation and risk controls.

What are common anti-patterns to avoid when designing Claude Code skills?

Common anti-patterns include inconsistent naming conventions, missing frontmatter validation, and failing to use progressive disclosure. Avoiding these ensures reliable behavior and maintains high-quality skills with proper risk controls.

Do I need to use a router pattern for simple Claude Code skills?

The router pattern is primarily designed for multi-workflow skills requiring intake and routing. Simple skills may not need this structure, but applying naming conventions and frontmatter validation remains essential for consistent discovery.