mcaf-skill-curation

Validate YAML frontmatter and folder names across .codex/skills and .claude/skills.

20|Updated Feb 1, 2026

Search Tags:#automation #workflow #codex #skills #validate #yaml-frontmatter #repo-management

Ready to run

npx skills add https://github.com/Zendevve/opentweak --skill mcaf-skill-curation-zendevve

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: mcaf-skill-curation
Source: https://github.com/Zendevve/opentweak/tree/main/.codex/skills/mcaf-skill-curation
Command: npx skills add https://github.com/Zendevve/opentweak --skill mcaf-skill-curation-zendevve

SYSTEM DOCUMENTATION & REQUIREMENTS

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

## What problem does it solve? This Skill helps teams keep agent skills in sync with the real codebase and governance rules by auditing, aligning frontmatter and folder structures across .codex/skills and .claude/skills, and by applying feedback to ensure skills match the repository reality.

## Core Features & Use Cases

Inventory and drift detection across skill folders to identify mismatches between SKILL.md frontmatter and actual folder names.
Validate and enforce YAML frontmatter rules (name, description, proper lengths, and folder-name alignment); remove template cruft (README.md, CHANGELOG.md, etc.) to keep skill folders lean.
Generate a metadata-only <available_skills> block and a validation report to help agent runtimes discover current capabilities.

Quick Start

Run the mcaf-skill-curation workflow to validate and generate available_skills metadata for your agent.

Frequently Asked Questions about mcaf-skill-curation

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

FAQPage Schema

How do I detect drift between agent skills and my repository definitions?▼

To detect agent skill drift, inventory and compare SKILL.md frontmatter against actual folder names across .codex/skills and .claude/skills to identify structural mismatches. This process surfaces inconsistencies between documented metadata and real codebase definitions.

How do I validate YAML frontmatter rules for agent skills?▼

To validate YAML frontmatter, enforce rules ensuring the name and description fields follow required formats, proper lengths, and folder-name alignment. This removes template cruft like README.md and keeps skill folders lean.

What is the best way to generate available skills metadata for agent runtimes?▼

The best way to generate available skills metadata is to run a curation workflow that audits skill folders and produces a metadata-only <available_skills> block plus a validation report. This helps agent runtimes discover current capabilities.

Can I use YAML validation across both .codex/skills and .claude/skills directories?▼

Yes, you can apply YAML validation across both .codex/skills and .claude/skills directories. The workflow audits both paths to align frontmatter rules, enforce folder-name matching, and remove template docs consistently.

Why does my agent skill folder name not match the frontmatter name?▼

Your agent skill folder name does not match the frontmatter name due to repository drift over time. Running a skill curation validation enforces folder-name alignment with the YAML frontmatter to resolve these mismatches.

Do I need the yaml dependency to validate agent skills?▼

Yes, you need the yaml dependency to parse and validate SKILL.md frontmatter rules. It enables the enforcement of proper name and description lengths and folder-name matching across skill directories.