create-skill

Create Codex skills from idea to deployment with YAML frontmatter and evaluation loops.

Updated Aug 23, 2026
One-click install
npx skills add https://github.com/lc0rp/luke-agent-scripts --skill create-skill-lc0rp
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: create-skill
Source: https://github.com/lc0rp/luke-agent-scripts/tree/main/skills/create-skill
Command: npx skills add https://github.com/lc0rp/luke-agent-scripts --skill create-skill-lc0rp

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires pyyaml, and includes scripts (resource) components.

What problem does it solve?

This Skill helps teams design Codex skills end-to-end—from capturing user intent and drafting SKILL.md with YAML frontmatter, to organizing optional resources (scripts, references, assets), running evaluation loops, analyzing results, and iterating toward a polished, well-triggered skill.

Core Features & Use Cases

  • Capture Intent: Define what the skill enables Codex to do and when it should trigger.
  • Draft SKILL.md: Create the YAML frontmatter (name, description, compatibility) and compelling Markdown body with operational instructions.
  • Bundle Resources: Organize optional scripts, references, and assets for on-demand use.
  • Run Evaluations: Execute trigger-evaluation loops, collect results, and generate structured evals.json and benchmark data.
  • Optimize Trigger Descriptions: Use the built-in improve-description flow to refine the skill's triggering description for better discovery and reliability.

Quick Start

Start by drafting a high-level skill concept and run the end-to-end evaluation loop to iterate toward a stable skill.

Frequently Asked Questions about create-skill

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create a Codex skill from scratch using YAML frontmatter?

Creating a Codex skill involves capturing user intent, drafting a SKILL.md file with YAML frontmatter for metadata, bundling optional scripts or references, and iterating through trigger-evaluation loops until the skill description is stable and reliably triggers.

What is a trigger-eval loop for skill evaluation and optimization?

A trigger-eval loop is an iterative process that executes skill trigger evaluations, collects structured results into evals.json, and analyzes benchmark data to refine the skill's triggering description for better discovery and reliable activation within Codex.

Do I need pyyaml installed to author and generate Codex skills?

Yes, pyyaml is a required dependency for this skill creation workflow. You need the pyyaml package installed in your environment to properly parse, generate, and manage the YAML frontmatter embedded in the SKILL.md files.

How do I optimize a Codex skill description for better trigger reliability?

Optimizing a skill description uses a built-in improve-description flow that analyzes evaluation results and iteratively refines the YAML frontmatter description field, ensuring the skill triggers correctly and is easily discovered by the Codex system.

Can I bundle external scripts and reference assets into a Codex skill?

Yes, you can organize and bundle optional resources including scripts, references, and assets alongside the SKILL.md file. These components are structured for on-demand use by Codex when the skill is triggered during execution.

What's the best way to structure SKILL.md for Codex skill automation?

The best approach is structuring SKILL.md with YAML frontmatter containing the skill name, description, and compatibility metadata, followed by a Markdown body with compelling operational instructions that guide the automation behavior clearly.