eval-skills

Audit SKILL.md files for frontmatter completeness and effort-level appropriateness.

Updated Apr 12, 2026
One-click install
npx skills add https://github.com/Alamator/code-ultimate --skill eval-skills
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: eval-skills
Source: https://github.com/Alamator/code-ultimate/tree/main/examples/skills/eval-skills
Command: npx skills add https://github.com/Alamator/code-ultimate --skill eval-skills

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill automates discovery and quality auditing of SKILL.md and flat skill markdown files across a repository to identify missing frontmatter, improper allowed-tools scoping, inappropriate effort tags, and structural/content issues. It consolidates findings into a scored report and produces recommended effort-field patches so maintainers can quickly remediate inconsistencies.

Core Features & Use Cases

  • Automated discovery of SKILL.md and flat .md skill files using filesystem globs and optional paths.
  • Structured scoring across name, description, allowed-tools, effort, content structure, and bonus metadata with clear thresholds for pass/fail.
  • Effort inference engine that suggests low/medium/high effort based on verbs, scope, allowed-tools, and content signals.
  • Report generation with per-skill tables, inferred effort, priority fixes, and a copy-paste-ready patch block for missing or mismatched effort fields.
  • Use Cases: onboarding a new project, pre-commit quality checks, bulk-import verification, and adding effort metadata to an existing skill library.

Quick Start

Run an audit of all SKILL.md files in this repository and output a scored report with recommended effort fields.

Frequently Asked Questions about eval-skills

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

FAQPage Schema
How do I audit and score skill markdown files for frontmatter completeness?

To audit and score skill markdown files, this skill parses YAML frontmatter, checks allowed-tools scoping, evaluates content structure, and generates a per-skill scored report with prioritized fixes.

How do I infer the correct effort level for a SKILL.md file?

Inferring the correct effort level for a SKILL.md file is done by analyzing verbs, scope, allowed-tools, and content signals to output a copy-paste-ready patch with recommended low, medium, or high effort fields.

What is the best way to automate quality assurance checks for skill files before a commit?

Automating quality assurance checks for skill files before a commit requires discovering files via filesystem globs, validating frontmatter metadata, and outputting a structured report identifying missing fields and structural issues.

Can I use this skill to bulk verify imported skill markdown files?

Yes, you can bulk verify imported skill markdown files by applying the audit to provided paths, which parses YAML frontmatter and generates per-skill scores and inferred effort levels across the entire import batch.

Do I need file system read access to discover and evaluate skill files?

Yes, file system read access, globbing capabilities, and simple shell execution are required to discover files, parse YAML frontmatter, analyze content structure, and produce recommended effort field patches.

Why does my skill audit report fail on allowed-tools scoping?

A skill audit report fails on allowed-tools scoping when the parsed YAML frontmatter contains missing, overly broad, or improperly defined tool permissions that do not meet the structured scoring thresholds.