skills.rest
mgajewskik
mgajewskikCOMMUNITY
Explore all skills by this author →
GitHub

skill-creator

Guide users through drafting, validating, and iterating Claude skills with SKILL.md and evals.json.

6|1|Updated Dec 29, 2025
Search Tags:#automation#evaluation#iteration#benchmark#SKILL.md#skill-creation#evals.json
One-click install
npx skills add https://github.com/mgajewskik/opencode-config --skill skill-creator-mgajewskik
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/mgajewskik/opencode-config/tree/main/skills/skill-creator
Command: npx skills add https://github.com/mgajewskik/opencode-config --skill skill-creator-mgajewskik

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

Creating consistent, high-quality Claude skills is hard: you must capture intent, write structured SKILL.md, design repeatable evals, and iterate based on results. This Skill guides you through drafting, validating, and improving skills end-to-end to reduce time-to-value and improve trigger accuracy.

Core Features & Use Cases

  • Define a skill by writing a formal SKILL.md frontmatter (name and description required) and an instructional body that Claude can load when activated.
  • Build and manage evals: generate eval prompts, specify expectations, and run iterative tests with and without the skill to measure impact.
  • Automate improvement loops: capture feedback, generate improved descriptions, and re-run evaluations to converge on a robust skill.

Quick Start

Draft a new skill by writing SKILL.md, configure an eval set, and run the evaluation loop to begin iterative improvement.

Frequently Asked Questions about skill-creator

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

FAQPage Schema
How do I create a Claude skill with automated evaluation?▼

Creating a Claude skill involves writing a SKILL.md file with valid name and description frontmatter, then running repeatable evaluation loops to validate and iterate on trigger accuracy.

What is the best way to benchmark skill trigger accuracy?▼

Benchmarking skill trigger accuracy requires generating eval prompts, specifying expectations, and running iterative tests with and without the skill to measure behavioral impact.

How does iterative skill improvement work with evals.json?▼

Iterative skill improvement uses evals.json to organize evaluation prompts, capture feedback and timing, generate improved descriptions, and re-run evaluations to converge on a robust skill.

Do I need PyYAML to draft and validate SKILL.md frontmatter?▼

PyYAML is required as a dependency to parse and validate the YAML frontmatter structure needed for skill discovery and activation within the automation workflow.

Can I use bundled scripts and references on demand during skill creation?▼

Bundled scripts, references, and assets are available on demand during skill creation to support drafting, validation, and evaluation tasks without requiring external setup.

Why does my skill discovery require valid frontmatter?▼

Skill discovery requires valid name and description frontmatter because the system reads these fields to identify when the skill should be activated and loaded by Claude.

Related Skills

Explore More Software Engineering Skills→
tuliosouza99tuliosouza99

skill-creator

Create and iteratively improve SKILL.md-based agent skills with Python eval scripts.

Community
Advanced
AnveioAnveio
1

kpi-verification-pass-rate

Measure first-pass verification rates for automated code verification pipelines.

Community
Intermediate
ulpi-ioulpi-io
2

writing-tools

Enforce a red-green-refactor workflow for creating and editing skill documentation.

Official
Advanced
phoenicisfphoenicisf

skill-creator

Generate, validate, and refine SKILL.md definitions for Claude skills.

Community
Advanced