create-agent-skills

Author Claude Skills with SKILL.md structure and router-pattern guidance.

1|Updated Jan 19, 2026
One-click install
npx skills add https://github.com/cercova-studios/terminal-agent-plugins --skill create-agent-skills-cercova-studios
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: create-agent-skills
Source: https://github.com/cercova-studios/terminal-agent-plugins/tree/main/plugins/10x-swe/skills/create-agent-skills
Command: npx skills add https://github.com/cercova-studios/terminal-agent-plugins --skill create-agent-skills-cercova-studios

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill helps teams design, implement, and maintain modular Claude Skill units with consistent structure, inline essential principles, and progressive disclosure to reduce cognitive load and token usage.

Core Features & Use Cases

  • Router-pattern guidance: split SKILL.md, workflows/, references/, templates/, and scripts/ for scalable skills.
  • Inline essential principles: ensure non-negotiable patterns are loaded with every invocation.
  • Progressive disclosure: load only needed references and workflows to minimize token usage.
  • Verification & auditing: built-in workflows for auditing, upgrading, and verifying skill health.
  • Domain knowledge reuse: references repository to share patterns, anti-patterns, and best practices across skills.
  • Easy growth: simple skill (single file) or complex router pattern with multiple workflows.

Quick Start

Create a simple one-file skill by placing a SKILL.md at the root with YAML frontmatter, and include a minimal <objective>, <quick_start>, and <success_criteria>.

Frequently Asked Questions about create-agent-skills

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

FAQPage Schema
How do I structure a Claude Skill to reduce token usage and cognitive load?

To structure a Claude Skill and reduce token usage, apply progressive disclosure by loading only needed references and workflows. This approach uses an XML-first structure to minimize cognitive load and optimize token consumption during skill invocation.

What is the best way to build a router-pattern skill with multiple workflows?

The best way to build a router-pattern skill is to split your project into SKILL.md, workflows/, references/, templates/, and scripts/ directories. This modular structure enables scalable skill design and consistent quality across complex implementations.

How do I create a simple one-file Claude Skill with XML?

Create a simple one-file Claude Skill by placing a SKILL.md file at the root with YAML frontmatter. Include minimal XML components like <objective>, <quick_start>, and <success_criteria> to define the skill's core functionality.

Can I use a single SKILL.md file for simple tasks instead of a router pattern?

Yes, you can use a single SKILL.md file for simple tasks instead of a router pattern. The skill framework supports both simple single-file implementations and complex router patterns with multiple workflows, allowing easy growth as complexity increases.

How do I verify and audit the health of Claude Skills?

Verify and audit the health of Claude Skills using built-in workflows for auditing, upgrading, and verifying skill health. These workflows enforce clear verification standards and ensure consistent quality across your skill repository.

When should I split skill domain knowledge into separate references?

Split domain knowledge into separate references when you need to share patterns, anti-patterns, and best practices across multiple skills. This progressive disclosure approach loads only needed references, reducing token usage and cognitive load during invocation.