skill-creator

Draft, validate, and package Claude Skills into .skill files.

Updated Feb 11, 2016
One-click install
npx skills add https://github.com/Ehrax/dotfiles --skill skill-creator-ehrax
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/Ehrax/dotfiles/tree/main/configs/agents/skills/skill-creator
Command: npx skills add https://github.com/Ehrax/dotfiles --skill skill-creator-ehrax

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

Creating and maintaining Claude skills is a complex, multi-step workflow. This Skill unit standardizes the end-to-end process—from drafting SKILL.md frontmatter, writing the instructional body, to running iterative evals, benchmarking results, and packaging deployable skills—so you can consistently produce high-quality skill units.

Core Features & Use Cases

  • Draft and validate SKILL.md frontmatter and body, including optional resources (scripts/, references/, assets/)
  • Run trigger-evaluation loops to measure triggering accuracy, in-context guidance, and reliability
  • Benchmark results across configurations, analyze variance, and identify improvements
  • Iterate on descriptions and skills, then package into a .skill file for deployment
  • Generate and share evaluation artifacts, including evals.json, benchmark.json, and viewer reports

Quick Start

To begin, draft a SKILL.md, run the evaluation loop, iterate on feedback, and package the skill into a .skill file.

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 Claude skills from scratch?

Creating and validating Claude skills involves drafting the SKILL.md frontmatter and body, enforcing a strict metadata schema, bundling optional resources like scripts, and running validation loops to ensure a deployable unit.

What is a trigger-evaluation loop for benchmarking skills?

A trigger-evaluation loop measures a skill's triggering accuracy, in-context guidance, and reliability. It runs iterative evals, analyzes variance across configurations, and identifies improvements to refine the skill before packaging.

Do I need yaml to package a deployable Claude skill?

Yes, yaml is a required dependency to package a deployable Claude skill. It enforces the strict metadata schema validation needed to properly format the SKILL.md frontmatter before bundling into a .skill file.

How do I package a Claude skill into a deployable unit?

Packaging a Claude skill into a deployable unit requires validating the SKILL.md metadata, bundling optional resources such as scripts, references, and assets, and exporting the finalized structure into a .skill file for deployment.

What evaluation artifacts does skill benchmarking generate?

Skill benchmarking generates evaluation artifacts including evals.json, benchmark.json, and viewer reports. These artifacts capture triggering accuracy, in-context guidance metrics, and variance analysis across configurations.

Can I bundle scripts and references when creating Claude skills?

Yes, you can bundle optional resources including scripts, references, and assets when creating Claude skills. These components are validated alongside the SKILL.md frontmatter and body during the packaging process.