skill-creator

Design, validate, and deploy Claude skills with SKILL.md frontmatter and evaluation loops.

Updated Mar 9, 2026
One-click install
npx skills add https://github.com/whchi/agentic-coding --skill skill-creator-whchi
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/whchi/agentic-coding/tree/main/skills/skill-creator
Command: npx skills add https://github.com/whchi/agentic-coding --skill skill-creator-whchi

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires PyYAML, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

The Skill provides a structured framework to create, document, and improve Claude skills, enabling teams to design, test, and deploy reliable capabilities with measurable evaluation loops.

Core Features & Use Cases

  • Define a new skill with a SKILL.md frontmatter and detailed instructions that guide activation and usage.
  • Run iterative evals, compare configurations (with_skill vs without_skill), and benchmark performance to drive continuous improvement.
  • Package skills as distributable .skill files, including optional resources (scripts, references, assets) for producer-grade deployment.

Quick Start

Create a new skill directory with SKILL.md (name and description), add optional resources, then run the evaluation loop to start refinement.

Frequently Asked Questions about skill-creator

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

FAQPage Schema
How do I create and validate a Claude skill with YAML frontmatter?

To create a Claude skill, you define a SKILL.md file with YAML frontmatter and body instructions, then run quick_validate to check frontmatter constraints and safety. This framework ensures your skill is structured correctly before deployment.

What is the best way to evaluate and benchmark Claude skill performance?

Evaluating Claude skill performance is done by running iterative eval loops that compare configurations with_skill vs without_skill. This benchmark comparison process drives continuous improvement and helps refine skill instructions based on measurable results.

Can I package Claude skills as distributable files with scripts and assets?

Yes, you can package Claude skills as distributable .skill files that include optional resources like scripts, references, and assets. This producer-grade packaging enables reliable deployment across teams and multiple domains.

Do I need PyYAML to build and test Claude skills?

Yes, PyYAML is required as a dependency to build and test Claude skills. It handles parsing the YAML frontmatter in your SKILL.md file, which is essential for the validation and description optimization processes.

How does the grader tool improve skill description optimization?

The grader tool improves skill description optimization by applying safety checks and evaluation loops to your SKILL.md file. It assesses the frontmatter and body instructions to ensure reliable activation and measurable performance.