skill-creator

Automate creation, modification, evaluation, and benchmarking of Claude skills.

15|2|Updated Feb 12, 2026
One-click install
npx skills add https://github.com/kousukef/forge --skill skill-creator-kousukef
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/kousukef/forge/tree/main/skills/skill-creator
Command: npx skills add https://github.com/kousukef/forge --skill skill-creator-kousukef

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

The Skill Creator provides a structured, repeatable workflow to turn user needs into fully documented Claude skills. It guides you from conceiving a skill, drafting SKILL.md, writing tests, and iterating based on eval results to ensure scalable skill development and stable triggering accuracy.

Core Features & Use Cases

  • End-to-end skill drafting: capture intent, draft SKILL.md, and assemble optional resources (scripts, references, assets).
  • Evaluation and benchmarking: design evals, run iterative tests, and analyze results to improve performance.
  • Description optimization: refine triggering text to maximize activation accuracy with real users.
  • Collaboration and governance: iterate skills with feedback, versioning, and reusability across projects.

Quick Start

Describe your idea in one sentence and I will guide you through drafting the SKILL.md and setting up the first eval.

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 a repeatable workflow from draft to production?

To create a Claude skill, you draft a SKILL.md with required frontmatter, assemble optional scripts, references, and assets directories, then run iterative evaluations to benchmark performance and refine the triggering description for production.

What is a SKILL.md file and how does it structure Claude skill creation?

A SKILL.md file is the core document for Claude skill creation that uses frontmatter to define metadata for accurate triggering, while optional scripts, references, and assets directories provide extended functionality for the skill.

How do I write and run evals to benchmark Claude skill performance?

You benchmark Claude skill performance by designing evaluation tests, running iterative evals against the drafted SKILL.md, and analyzing the resulting evaluation artifacts to track performance and improve triggering accuracy.

Can I optimize skill description text to improve activation accuracy with real users?

You can optimize skill description text by refining the triggering frontmatter based on iterative evaluation results, which maximizes activation accuracy and ensures the skill triggers reliably for real user queries.

What is the best way to manage skill versioning and governance across multiple projects?

The best way to manage skill versioning and governance is to iterate skills with structured feedback, maintain evaluation artifacts for performance tracking, and ensure reusability across projects through a repeatable draft-to-production workflow.

Do I need to set up scripts and references directories before drafting a new skill?

You do not need to set up scripts and references directories initially, as they are optional for extended functionality; you only need to maintain the core SKILL.md with frontmatter to begin the skill creation workflow.