skills.rest
pferretti99
pferretti99COMMUNITY
Explore all skills by this author →
GitHub

skill-creator

Create or refine Claude skill directories with SKILL.md and evaluation loops.

1|Updated Mar 13, 2026
Search Tags:#prompt engineering#skill creation#benchmarking#evaluation#iterative improvement#trigger optimization
One-click install
npx skills add https://github.com/pferretti99/learnshareremix-2026-labcamp-cursor-vs-copilot --skill skill-creator-pferretti99
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/pferretti99/learnshareremix-2026-labcamp-cursor-vs-copilot/tree/main/skills/anthropic-official/skills/skill-creator
Command: npx skills add https://github.com/pferretti99/learnshareremix-2026-labcamp-cursor-vs-copilot --skill skill-creator-pferretti99

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill streamlines the process of creating a new Claude coding skill (or improving an existing one) by turning your idea into a working SKILL.md, defining a test set, running trigger evaluations, and iterating based on both qualitative and quantitative results.

Core Features & Use Cases

  • Draft skills from your intent: Convert what you want a skill to do into a first working version of SKILL.md.
  • Design and run evals: Produce realistic test prompts, run trigger checks, and quantify results (including variance/benchmarking where applicable).
  • Iterate until it performs: Use evaluation feedback to repeatedly rewrite and optimize the skill’s description and behavior.
  • Improve triggering accuracy: Optionally run a description-optimization loop to increase correct skill invocation rates.

Quick Start

Ask this skill to create a new skill from your idea, then tell it the kinds of user prompts and outputs you expect so it can draft SKILL.md and propose an eval set.

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 new Claude skill from scratch?▼

To create a new Claude skill, you provide your intent and expected prompts, and the system drafts a SKILL.md file with YAML frontmatter, then generates test sets to evaluate and iterate on performance.

What is trigger optimization for Claude skills?▼

Trigger optimization is an automated loop that rewrites a skill's description to increase correct invocation rates, using baseline evaluations and variance analysis to measure improvements.

How do I benchmark an existing skill to test improvements?▼

You benchmark an existing skill by running structured evaluation prompt sets against a baseline, then applying variance analysis to quantify measurable performance differences after changes.

Can I iteratively improve skill instructions and triggering together?▼

Yes, you can iteratively improve both instructions and triggering by running evaluation loops that use qualitative and quantitative feedback to repeatedly rewrite and optimize the skill.

Do I need YAML frontmatter for Claude skill creation?▼

Yes, YAML frontmatter containing the skill name and description is required, as it defines the metadata used for triggering and structured evaluation during the creation process.

Related Skills

Explore More Software Engineering Skills→
syahiidkamilsyahiidkamil
316

skill-creator

Guide users through designing, documenting, and evaluating Claude Code skills.

Community
Advanced
ArogyaReddyArogyaReddy

llm-evaluation

Automates LLM output evaluation with metrics, human feedback, and benchmarking.

Community
Advanced
OGismatulinOGismatulin

skill-creator

Develop, test, and optimize AI prompt skills through iterative evaluation cycles.

Community
Intermediate
cekura-aicekura-ai
5

Cekura Metric Design

Creates custom AI and machine-learning system evaluation metrics and workflows.

Official
Advanced