skill-creator

Create, modify, and evaluate Claude Skills with SKILL.md files.

Updated Apr 8, 2026
One-click install
npx skills add https://github.com/ThuHuynh1438/pm-kit --skill skill-creator-thuhuynh1438
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/ThuHuynh1438/pm-kit/tree/main/.claude/skills/skill-creator
Command: npx skills add https://github.com/ThuHuynh1438/pm-kit --skill skill-creator-thuhuynh1438

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires anthropic, pyyaml, and includes scripts (resource) components.

What problem does it solve?

The Skill helps teams quickly create new Claude skills, iteratively improve existing ones, and measure skill performance with structured evaluations.

Core Features & Use Cases

  • End-to-end skill creation workflow: draft SKILL.md, wire in references/scripts, and prepare evaluation prompts.
  • Iterative improvement and benchmarking: run eval loops, analyze results, and generate improvements to trigger phrases.
  • Descriptive guidance and governance: provide processes for evaluating skill quality and maintaining versioned iterations.

Quick Start

Draft a new skill by outlining the concept, writing the initial SKILL.md, then run the evaluation loop to test and refine the skill.

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 evaluate Claude skills from scratch?

To create and evaluate Claude skills, draft an initial SKILL.md from your concept, wire in optional resources like scripts or references, then run the evaluation loop to test and refine the skill.

What is the best way to benchmark prompt performance for custom workflows?

Benchmarking prompt performance requires running structured evaluation loops to analyze results and generate improvements to trigger phrases, ensuring your custom workflows achieve production-ready quality.

Can I use Python scripts to automate skill iteration and frontmatter validation?

Yes, you can automate skill iteration and frontmatter validation using Python scripts with dependencies like anthropic and pyyaml to ensure formalized flow and structured governance.

How does iterative improvement work for refining existing Claude skills?

Iterative improvement works by running evaluation loops on existing skills, analyzing the benchmarked results, and generating targeted refinements to trigger phrases and supporting artifacts.

Do I need specific dependencies to package production-ready Claude skills?

Yes, packaging production-ready Claude skills requires specific dependencies including anthropic and pyyaml to support frontmatter validation, end-to-end tooling, and automated evaluation workflows.

What are the limitations when drafting SKILL.md files for complex prompt-engineering workflows?

When drafting SKILL.md files for complex prompt-engineering workflows, limitations may arise if the evaluation loops cannot adequately benchmark highly nuanced trigger phrases, requiring manual governance adjustments.