skill-creator

Draft, evaluate, and refine Claude Skills with SKILL.md and eval cycles.

Updated Feb 20, 2026
One-click install
npx skills add https://github.com/saajunaid/junai --skill skill-creator-saajunaid
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/saajunaid/junai/tree/main/.github/skills/workflow/skill-creator
Command: npx skills add https://github.com/saajunaid/junai --skill skill-creator-saajunaid

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires yaml.

What problem does it solve?

The Skill Creator guides users from an initial idea to a fully defined Skill Unit by capturing intent, drafting a robust SKILL.md, and orchestrating iterative evaluation and benchmarking workflows to improve triggering accuracy and reliability.

Core Features & Use Cases

  • Intent capture and SKILL.md drafting to articulate a clear skill identity
  • Lightweight eval/benchmark loops to validate triggering and performance
  • Automatic history tracking and iterative improvements to converge on a strong description and prompts
  • Reusable templates and patterns that speed up the creation and refinement of new skills

Quick Start

Draft an initial SKILL.md for your idea and start the intent-capture flow to trigger evaluation and iteration immediately.

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 refine Claude skills from an initial idea?

To create Claude skills, you capture user intent and draft a SKILL.md file, then run iterative evaluation and benchmarking loops to refine triggering accuracy and output quality until the skill is polished.

What is a SKILL.md file and why is frontmatter important for skill triggering?

A SKILL.md file defines a skill's identity and processing logic, using YAML frontmatter to store metadata that controls how accurately and reliably the skill triggers during workflows.

How do I set up an evaluation and benchmarking workflow for Claude skills?

You set up an evaluation workflow by generating specific test prompts, running lightweight benchmark loops to validate performance, and tracking versioned history to apply iterative improvements.

Do I need YAML to draft and manage skill metadata?

Yes, YAML is a required dependency for managing skill metadata, allowing you to structure frontmatter and configuration data needed to define and deploy robust skill units.

What is the best way to optimize skill descriptions for accurate triggering?

The best way to optimize descriptions is through automated iteration, using benchmark results to automatically orchestrate workflow revisions that converge on a strong, reliable triggering description.

When should I use an automated skill creation workflow instead of manual drafting?

Use an automated skill creation workflow when you need structured metadata, versioned history tracking, and iterative eval loops to converge on robust performance that manual drafting cannot easily validate.