agentleFS
mastepanoskimastepanoski/claude-skills/CLAUDE.md

claude-skills

CLAUDE.md62 starsChanged 9 months agoMITView on GitHub

This is an Agent Skills repository following the open standard that makes skills compatible with Claude Code, OpenAI Codex, ChatGPT, and other AI coding assistants. It contains UX/UI evaluation skills for auditing interfaces using industry-standard methodologies. Validate structure (descriptions ≤200 chars, frontmatter, naming, SKILL.md size) with the bundled script: To exercise a real install of your local working copy (including uncommitted branch changes), install into a scratch target outside the repo, passing the repo path as the source: ⚠️ Do…

The file

Show its rules and text

We split this CLAUDE.md into the rules it gives an agent

13 rules. About 0 are common, 8 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈8

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    After installing, invoke the skill through natural language (the AI agent will automatically detect when to use it based on the YAML …

    line 32
  2. Very specific:

    Every SKILL.md must follow this exact structure:

    line 49
  3. Very specific:

    Use kebab-case for skill names (e.g., wcag-accessibility-audit)

    line 75
  4. Distinctive:

    Update README.md: Add skill to the skills table

    line 86
  5. Very specific:

    Update CHANGELOG.md: Document the addition under [Unreleased]

    line 88
  6. Very specific:

    Scope: Use skill name or affected area (e.g., wcag-audit, contributing)

    line 122
  7. Distinctive:

    No capitalization on first letter

    line 126
  8. Distinctive:

    No period at end

    line 127
  9. Distinctive:

    Before committing a new skill, verify:

    line 165
  10. Distinctive:

    When modifying skills or documentation:

    line 178
  11. Distinctive:

    README.md: Update skills table, installation examples, or feature descriptions

    line 180
  12. Very specific:

    CHANGELOG.md: Add entry under [Unreleased] following Keep a Changelog format

    line 181
  13. Distinctive:

    Write skills to be agent-agnostic.

    line 192

≈ “Common” is an estimate: at least 1 in 5 of the 27,641 CLAUDE.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
CLAUDE.md

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Repository Purpose

This is an Agent Skills repository following the open standard that makes skills compatible with Claude Code, OpenAI Codex, ChatGPT, and other AI coding assistants. It contains UX/UI evaluation skills for auditing interfaces using industry-standard methodologies.

Testing Skills Locally

Validate structure (descriptions ≤200 chars, frontmatter, naming, SKILL.md size) with the bundled script:

npm run validate          # or: node scripts/validate.js

To exercise a real install of your local working copy (including uncommitted branch changes), install into a scratch target outside the repo, passing the repo path as the source:

mkdir -p /tmp/skills-test && cd /tmp/skills-test
npx skills add /absolute/path/to/this/repo --skill skill-name
# installed under /tmp/skills-test/.agents/skills/<skill-name>/ (SKILL.md + references/)

⚠️ Do not run npx skills add . from the repository root. The CLI treats the cwd as the install target, so it creates .agents/ inside the repo, replaces skills/<name>/ with a symlink into it, and writes skills-lock.json — clobbering the source. Keeping the target in /tmp (cwd ≠ repo) leaves the repo untouched; the repo is only read as the source.

After installing, invoke the skill through natural language (the AI agent will automatically detect when to use it based on the YAML description).

Repository Architecture

Core Structure

skills/                    # All skills live here
├── skill-name/           # Each skill is a directory
│   ├── SKILL.md          # REQUIRED: YAML frontmatter + markdown instructions
│   ├── scripts/          # OPTIONAL: Executable helpers (.sh, .py, .js)
│   ├── references/       # OPTIONAL: Reference documentation
│   └── assets/           # OPTIONAL: Templates, examples

SKILL.md Anatomy

Every SKILL.md must follow this exact structure:

---
name: skill-name                    # kebab-case, matches directory name
description: Brief description...   # MAX 200 chars - AI uses this to determine invocation
---

# Skill Title

## When to Use This Skill
[Scenarios for invocation]

## Inputs Required
[Required and optional parameters]

## Procedure
[Step-by-step instructions the AI agent executes]

## Output Format
[Expected output structure]

Critical constraints:

  • description field MUST be ≤200 characters (AI uses this for skill matching)
  • YAML frontmatter must parse correctly
  • Use kebab-case for skill names (e.g., wcag-accessibility-audit)

Adding New Skills

  1. Create directory structure:
mkdir -p skills/your-skill-name
  1. Create SKILL.md with YAML frontmatter + instructions
  2. Update README.md: Add skill to the skills table
  3. Update CHANGELOG.md: Document the addition under [Unreleased]
  4. Validate locally:
npm run validate

(For a real install test, run npx skills add from outside the repo — see "Testing Skills Locally".)

  1. Commit using Conventional Commits:
git commit -m "feat(your-skill-name): add skill for X evaluation"

Commit Convention (REQUIRED)

This repository uses Conventional Commits specification:

<type>(<scope>): <subject>

<body>

<footer>

Types:

  • feat: New skill or feature (feat(wcag-audit): add contrast checker)
  • fix: Bug fix (fix(nielsen): correct severity rating)
  • docs: Documentation (docs(readme): update examples)
  • style: Formatting only (style(skill): fix markdown)
  • refactor: Code restructuring (refactor(wcag): simplify structure)
  • test: Tests (test(skills): add validation)
  • chore: Maintenance (chore: update dependencies)

Scope: Use skill name or affected area (e.g., wcag-audit, contributing)

Subject rules:

  • Imperative mood: "add" not "added" or "adds"
  • No capitalization on first letter
  • No period at end
  • ≤50 characters

Examples:

# Good
feat(cognitive-walkthrough): add task analysis skill
fix(wcag-audit): correct WCAG 2.2 criteria references
docs(contributing): clarify testing process

# Bad
feat: new skill                          # Too vague, no scope
fix: fixed bugs                          # Not imperative, vague
feat(wcag): Added comprehensive audit    # Capitalized, not imperative

Current Skills

UX/UI Evaluation Suite:

  1. ux-audit-rethink - Holistic UX audit (IxDF): 7 factors + 5 usability characteristics + 5 interaction dimensions + redesign proposals
  2. don-norman-principles-audit - 7 principles: discoverability, affordances, signifiers, feedback, mapping, constraints, conceptual models
  3. nielsen-heuristics-audit - 10 usability heuristics with severity ratings (0-4)
  4. wcag-accessibility-audit - WCAG 2.1/2.2 compliance (A/AA/AAA levels, 4 POUR principles)
  5. cognitive-walkthrough - Task-specific deep-dive using 4 cognitive questions to evaluate learnability
  6. ui-design-review - Visual design evaluation: typography, color, spacing, hierarchy, consistency, branding (10 dimensions)

AI Governance & Security: 7. iso-42001-ai-governance - AI governance (ISO 42001:2023): Risk management, ethics, security, transparency, regulatory compliance (EU AI Act, GDPR) 8. nist-ai-rmf - AI risk assessment (NIST AI RMF 1.0): Govern, Map, Measure, Manage functions for trustworthy AI 9. owasp-llm-top10 - Security audit for LLM/GenAI apps (OWASP Top 10 for LLM Apps 2025) 10. owasp-ai-testing - AI trustworthiness testing (OWASP AI Testing Guide v1): 32 test cases across 4 layers 11. gdpr-audit - Technical GDPR audit of code, plans, schemas, or IaC: processing map + article-cited findings (not legal advice) 12. ai-assessment-scale - Measure AI contribution to a project using the AI Assessment Scale (AIAS) 5-level framework 13. information-architecture-audit - Information architecture (Rosenfeld/Morville four systems, 8 principles): organization, labeling, navigation, search, card sorting, tree testing 14. iso-27090-ai-security - AI security assessment (ISO/IEC 27090:2026): Threat-based gap assessment across 7 domains (governance, supply chain/AIBOM, data poisoning, model protection, inputs/outputs, monitoring) with maturity scale and client-ready report

Skill Quality Standards

Before committing a new skill, verify:

  • YAML description ≤200 characters
  • Clear "When to Use This Skill" section
  • Step-by-step procedure with actionable instructions
  • Defined output format
  • No spelling/grammar errors
  • Validated locally with npm run validate (do not run npx skills add . from the repo root)
  • Added to README.md skills table
  • Added to CHANGELOG.md under [Unreleased]
  • Uses kebab-case naming

Documentation Updates

When modifying skills or documentation:

README.md: Update skills table, installation examples, or feature descriptions CHANGELOG.md: Add entry under [Unreleased] following Keep a Changelog format CONTRIBUTING.md: Reference for contribution guidelines, testing, and PR process

Multi-Agent Compatibility

Skills in this repository follow the Agent Skills open standard (December 2025) adopted by:

  • Anthropic Claude Code
  • OpenAI Codex CLI
  • ChatGPT
  • Any agent supporting the specification

Write skills to be agent-agnostic. Avoid references to specific AI agents in skill instructions.

Installation Distribution

Users install skills via:

# From GitHub (primary)
npx skills add mastepanoski/claude-skills --skill skill-name

# From npm (if published)
npx skills add @mastepanoski/claude-skills --skill skill-name

The files field in package.json ensures only skills/ directory is distributed.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.