skill-creator

Automate drafting, testing, and refining Claude skills with evaluation loops.

Updated Mar 19, 2026
One-click install
npx skills add https://github.com/japurcell/skills --skill skill-creator-japurcell
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/japurcell/skills/tree/main/skills/skill-creator
Command: npx skills add https://github.com/japurcell/skills --skill skill-creator-japurcell

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

Automates the end-to-end lifecycle of skill creation and improvement by orchestrating structured evaluation loops that draft, test, and refine Claude skills, enabling iterative quality improvements with history tracking and validation.

Core Features & Use Cases

  • Draft, test, and iterate on SKILL.md frontmatter and body to ensure triggering accuracy and correct behavior.
  • Run automated evaluations, generate reports, and compare skill variants across configurations to identify the best performing description and prompts.
  • Package skills into distributable .skill files for deployment, reuse, and sharing, with optional bundled resources like scripts, references, and assets.

Quick Start

Draft a new skill in SKILL.md and run the built-in loop to evaluate prompts, review results, and iteratively rewrite the skill description.

Frequently Asked Questions about skill-creator

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

FAQPage Schema
How do I automate Claude skill creation and prompt testing?

You can automate Claude skill creation by running structured evaluation loops that draft, test, and refine SKILL.md descriptions iteratively. The evaluation loop applies training and test splits to evaluate prompt behavior and track improvement history across variants.

What is the best way to compare Claude skill variants during development?

Comparing Claude skill variants requires running automated evaluations with optional benchmarks to generate performance reports. This surfaces the best performing prompts and descriptions by testing multiple configurations against each other.

How do I package Claude skills for distribution and reuse?

Packaging Claude skills into distributable .skill files involves validating SKILL.md frontmatter and bundling resources like scripts, references, and assets. This provides a distributable format for deployment, reuse, and sharing.

Does skill automation require dependencies to run evaluation loops?

Skill automation evaluation loops require the pyyaml dependency to function. You need this installed to parse configuration files and run the automated testing workflows that validate and refine skill variants.

Why does my SKILL.md frontmatter validation fail during skill packaging?

SKILL.md frontmatter validation fails when required metadata fields are missing or improperly formatted. The evaluation tooling checks frontmatter structure during the packaging workflow to ensure the skill triggers correctly before distribution.