skill-creator

Design, evaluate, and refine Claude skills with a repeatable workflow.

4|Updated Feb 23, 2026
One-click install
npx skills add https://github.com/fysoul17/one-query-v1 --skill skill-creator-fysoul17
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/fysoul17/one-query-v1/tree/main/.agents/skills/skill-creator
Command: npx skills add https://github.com/fysoul17/one-query-v1 --skill skill-creator-fysoul17

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

This Skill provides a repeatable framework for creating, refining, and validating Claude skills. It guides users from initial drafting of SKILL.md, through writing test prompts and running evaluation loops, to iterating on descriptions and packaging a deployable skill.

Core Features & Use Cases

  • End-to-end workflow for constructing new skills, improving existing ones, and measuring performance.
  • Automated evaluation loops: generate prompts, run trigger evaluations, and iterate based on results.
  • Packaging and distribution: validate, bundle, and publish a portable .skill artifact.

Quick Start

Draft your skill, then run the eval loop to test, iterate, and package a deployable 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 test Claude skills with an automated evaluation loop?

To create and test Claude skills, draft a SKILL.md file, then run an automated evaluation loop to generate test prompts, evaluate trigger conditions, and iterate on the skill descriptions based on the results.

What is the best way to package a deployable Claude skill artifact?

The best way to package a deployable Claude skill is to validate the drafted SKILL.md, bundle the required scripts and assets, and export a portable .skill artifact ready for distribution.

How does automated evaluation work when refining Claude skill descriptions?

Automated evaluation works by generating test prompts, running trigger evaluations against your drafted skill, and providing benchmark feedback so you can iterate on the descriptions and improve performance.

Do I need YAML to write and validate a SKILL.md file?

Yes, YAML is a required dependency for drafting and validating the SKILL.md file, ensuring the configuration and metadata meet the necessary structure before packaging the deployable artifact.

Can I use this workflow to improve existing Claude skills, or is it only for new ones?

You can use the workflow to improve existing Claude skills by running evaluation loops on current descriptions, benchmarking the results, and iterating on the text to refine trigger accuracy and overall performance.

Are there limitations when packaging Claude skills with scripts and reference assets?

Packaging requires all component dependencies, including scripts and reference assets, to be properly configured and validated before bundling, ensuring the final .skill artifact functions correctly without missing files.