skill-creator

Draft OpenClacky Skills and run evaluation loops with benchmark reports.

1.2k|100|Updated Dec 30, 2025

Search Tags:#automation #evaluation #creation #benchmark #skill #prompt-engineering #loop

Ready to run

npx skills add https://github.com/clacky-ai/openclacky --skill skill-creator-clacky-ai

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/clacky-ai/openclacky/tree/main/lib/clacky/default_skills/skill-creator
Command: npx skills add https://github.com/clacky-ai/openclacky --skill skill-creator-clacky-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

Quickly transform user ideas into fully defined OpenClacky Skills, and manage their evaluation loops from draft to deployment.

Core Features & Use Cases

Interactive workflow: draft SKILL.md, run trigger evaluations, and iterate toward a robust skill.
Evaluation loop: conduct train/test splits, refine descriptions, and generate benchmark reports.
Description optimization: feed results back to the Skill description to improve triggering accuracy.

Quick Start

Provide a high-level idea for a new Skill, and this tool will draft SKILL.md and start the iterative evaluation loop to optimize the description.

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 evaluate an OpenClacky Skill from a high-level idea?▼

To create an OpenClacky Skill, provide a high-level idea and the tool drafts a SKILL.md file, then runs an evaluation loop with train/test splits to validate and refine the skill.

What is an evaluation loop for skill creation and how does it work?▼

An evaluation loop conducts train/test splits on trigger prompts, refines skill descriptions, and generates benchmark reports to improve triggering accuracy and deterministic validation.

Do I need a repository to automate benchmark validation for my prompts?▼

Yes, automating benchmark validation requires a repository containing evaluation prompts and scripts to enable deterministic validation and iterative improvement of the skill.

How do I optimize prompt descriptions to improve skill triggering accuracy?▼

You optimize prompt descriptions by running trigger evaluations, feeding the benchmark results back into the skill description, and iterating until triggering accuracy improves.

What is the best way to refine a draft skill into a validated deployment?▼

The best way to refine a draft skill is using an interactive workflow that drafts the initial definition, runs trigger evaluations, and iterates toward a robust, validated deployment.

Can I generate benchmark reports for my prompt engineering workflows?▼

Yes, you can generate benchmark reports by executing the evaluation loop, which performs train/test splits and deterministic validation to measure and improve prompt engineering outcomes.