skill-creator

Create and validate Claude Skill drafts with SKILL.md trigger descriptions.

1|Updated Jan 7, 2026
One-click install
npx skills add https://github.com/adamatdevops/forge-works --skill skill-creator-adamatdevops
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/adamatdevops/forge-works/tree/main/.skills/anthropics/skill-creator
Command: npx skills add https://github.com/adamatdevops/forge-works --skill skill-creator-adamatdevops

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and assets (resource) components.

What problem does it solve?

This Skill guides you through designing a new Claude skill—or improving an existing one—by turning your intent into a working draft, validating it with realistic test prompts, and iterating based on measured results.

Core Features & Use Cases

  • Skill draft from intent: Convert what you want into a complete SKILL.md structure, including actionable triggering guidance in the description frontmatter.
  • Test-prompt + evaluation loop: Generate and run a test set, collect qualitative outputs, and drive iteration using quantitative trigger-evaluation results.
  • Description trigger optimization: Improve the description field so Claude invokes the skill at the right times without over-triggering.
  • Flexible workflow: Supports jumping into different stages depending on whether you’re starting from scratch or already have a draft.

Quick Start

Ask this Skill to create a new skill that helps Claude “turn messy meeting notes into a clean action-item list,” and request test prompts plus an evaluation plan to verify correct triggering and output quality.

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 from a prompt intent?

To create a Claude skill, convert your prompt intent into a precise SKILL.md file with actionable triggering guidance in the description frontmatter. This structures consistent tool use and reliable trigger behavior across diverse prompts.

How does trigger evaluation work for Claude skills?

Trigger evaluation works by generating test queries, running Claude trigger evaluations, grading outcomes, and improving description text under a hard character budget. This iterative process guards against overfitting and incorrect triggering.

What is the best way to optimize skill descriptions without over-triggering?

Optimizing skill descriptions requires iteratively validating trigger evaluations against generated test prompts. You refine the description text under a hard character budget to ensure Claude invokes the skill at the right times without over-triggering.

Can I iterate on an existing SKILL.md draft instead of starting from scratch?

You can iterate on an existing SKILL.md draft using a flexible workflow that supports jumping into different stages. This allows you to validate existing drafts with realistic test prompts and drive improvements based on measured trigger-evaluation results.

Why does my Claude skill fail to trigger consistently across diverse prompts?

Inconsistent triggering happens when the description frontmatter lacks precise guidance. You fix this by running test-prompt evaluation loops, collecting quantitative trigger-evaluation results, and iteratively improving the description text.

Do I need to provide test prompts to validate Claude skill creation?

You need test prompts to validate skill creation and ensure reliable triggering behavior. The skill creation workflow generates a test set, runs Claude trigger evaluations, and grades outcomes to drive iterative improvements.