skill-creator

Create and optimize Claude skills with SKILL.md files and test iterations.

2|1|Updated Apr 1, 2026
One-click install
npx skills add https://github.com/miriamdong/mstack --skill skill-creator-miriamdong
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/miriamdong/mstack/tree/main/skill-creator
Command: npx skills add https://github.com/miriamdong/mstack --skill skill-creator-miriamdong

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Drafting, evaluating, and iterating Claude skills is complex and time-consuming. This section helps teams quickly define the skill's purpose, align on workflow, and set up a repeatable improvement loop.

Core Features & Use Cases

  • Capture Intent: Define what the skill should enable and when it should trigger.
  • Interview & Research: Gather edge cases, draft test prompts, and research best practices to shape the skill.
  • Write the SKILL.md: Create the frontmatter, body, and optional bundled resources (scripts, references, assets) for a real user workflow.
  • Run Evals & Iterate: Build test sets, execute evaluations, analyze results, and iterate on the skill to improve triggering accuracy and usefulness.
  • Description Optimization: Optionally optimize the triggering description to improve discoverability and triggering rate.

Quick Start

Draft an initial SKILL.md with frontmatter and a brief workflow description that a user can adapt.

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 iterate on Claude skills?

To create Claude skills, draft a SKILL.md file with frontmatter and workflow descriptions, then run evaluations and iterate on the content to improve trigger accuracy and overall usefulness.

What is the best way to write a SKILL.md file for Claude prompt-engineering?

Writing a SKILL.md file involves defining frontmatter, a body section, and optional bundled resources like scripts or references to capture the intended workflow and trigger conditions for Claude.

How do I set up evaluations to improve my Claude skill triggering accuracy?

You can improve Claude skill triggering accuracy by building test sets, executing evaluations, analyzing the results, and iterating on the description to optimize discoverability and trigger rates.

Can I optimize the description of a Claude skill after it is already deployed?

You can optimize an existing Claude skill description by running targeted evaluations and iterating on the text specifically to improve discoverability and increase the accuracy of workflow triggers.

How do I capture edge cases when designing a new workflow for Claude?

Capturing edge cases for a Claude workflow requires conducting interviews, researching best practices, and drafting test prompts to ensure the skill handles specific user scenarios effectively.

Do I need any specific dependencies to test and optimize Claude skills?

No specific dependencies are required to test and optimize Claude skills, as the workflow focuses on drafting SKILL.md files, generating test sets, and iterating on descriptions natively.