# Install skills from marketplace
/plugin marketplace add anthropics/skills

# Install from local directory
/plugin add /path/to/skill-directory

# Install community skill collections
/plugin marketplace add obra/superpowers-marketplace

import anthropic
import os

client = anthropic.Client(api_key=os.environ.get("ANTHROPIC_API_KEY"))

# List available skills
response = client.skills.list()

# Use skills in conversation
message = client.messages.create(
    model="claude-3-5-sonnet-20241022",
    max_tokens=1024,
    messages=[
        {"role": "user", "content": "Create a Word document with quarterly report"}
    ],
    skills=["docx"]
)

my-skill/
├── SKILL.md          # Main skill file with frontmatter
├── scripts/          # Optional executable scripts
│   └── helper.py
└── resources/        # Optional supporting files
    └── template.json

---
name: my-custom-skill
description: Brief description for skill discovery (keep concise)
triggers:
  - "phrase users might say"
  - "another trigger phrase"
---

# Detailed Instructions

Claude will read these instructions when the skill is activated.

## Usage
Explain how to use this skill...

## Examples
Provide clear examples...

# In Claude Code or Claude.ai
User: "Use the skill-creator to help me build a skill for [your task]"

# Follow interactive prompts to generate skill structure

---
name: api-documentation-generator
description: Generate comprehensive API documentation from code
triggers:
  - "generate API docs"
  - "document my API endpoints"
  - "create API reference"
---

# API Documentation Generator

This skill helps create comprehensive API documentation from code.

## Usage

1. Provide the API code or endpoint definitions
2. Specify the documentation format (OpenAPI, Markdown, etc.)
3. Claude generates structured documentation

## Example

For a Flask API:

```python
from flask import Flask, jsonify, request

app = Flask(__name__)

@app.route('/api/users', methods=['GET'])
def get_users():
    """Retrieve all users"""
    return jsonify({"users": []})

@app.route('/api/users/<int:user_id>', methods=['GET'])
def get_user(user_id):
    """Retrieve a specific user by ID"""
    return jsonify({"user_id": user_id})

## Key Official Skills

### Document Manipulation

```python
# docx skill - Word document handling
# Example usage through Claude
"Create a Word document with the following sections: Introduction, Methods, Results"

# pdf skill - PDF operations
"Extract tables from this PDF and convert to Excel format"

# xlsx skill - Excel spreadsheet handling
"Create a spreadsheet with sales data and generate pivot tables"

# pptx skill - PowerPoint presentations
"Create a 10-slide presentation about quarterly results with charts"

# frontend-design skill
"Create a landing page with bold design choices, avoid generic AI aesthetics"

# web-artifacts-builder skill
"Build an interactive dashboard using React, Tailwind, and shadcn/ui"

# mcp-builder skill
"Create an MCP server that integrates with the GitHub API"

# webapp-testing skill
"Test my local webapp at http://localhost:3000 using Playwright"

# algorithmic-art skill
"Create generative art using p5.js with flow fields and particle systems"

# canvas-design skill
"Design a poster for a tech conference in PNG format"

# slack-gif-creator skill
"Create an animated GIF for Slack showing a progress bar animation"

# Install the comprehensive skills library
/plugin marketplace add obra/superpowers-marketplace

# Available commands after installation
/brainstorm          # Interactive brainstorming
/write-plan         # Create detailed plans
/execute-plan       # Execute planned tasks
/skills-search      # Search available skills

# iOS Simulator automation
/plugin add https://github.com/conorluddy/ios-simulator-skill

# Web fuzzing with ffuf
/plugin add https://github.com/jthack/ffuf_claude_skill

# Browser automation with Playwright
/plugin add https://github.com/lackeyjb/playwright-skill

# D3.js visualizations
/plugin add https://github.com/chrisvoncsefalvay/claude-d3js-skill

# Scientific computing
/plugin add https://github.com/K-Dense-AI/claude-scientific-skills

# Security auditing
/plugin add https://github.com/trailofbits/skills

# Expo app development
/plugin add https://github.com/expo/skills

# ✅ Good - Use environment variables
api_key: ${API_KEY}
database_url: ${DATABASE_URL}

# ❌ Bad - Never hardcode secrets
api_key: "sk-1234567890abcdef"

# Test locally before sharing
/plugin add /path/to/my-skill

# Verify activation
"Use my-skill to [trigger phrase]"

# Check skill loading
# Skills should load metadata first, then full instructions when relevant

{
  "skills": {
    "enabled": true,
    "paths": [
      "/path/to/custom/skills",
      "~/.claude/skills"
    ]
  }
}

{
  "skills": [
    "frontend-design",
    "webapp-testing",
    "custom-skill"
  ],
  "skill_paths": [
    "./skills"
  ]
}

---
name: python-testing-skill
description: Generate Python tests with pytest
triggers:
  - "write tests for"
  - "create pytest tests"
---

# Detect when to activate
This skill activates when:
- Python files are present
- User requests test creation
- pytest is mentioned or available

complex-skill/
├── SKILL.md
├── scripts/
│   ├── analyzer.py
│   ├── generator.py
│   └── validator.py
└── resources/
    ├── templates/
    │   ├── template1.json
    │   └── template2.json
    └── config.yaml

## Available Scripts

- `scripts/analyzer.py` - Analyzes input data
- `scripts/generator.py` - Generates output
- `scripts/validator.py` - Validates results

## Templates

Located in `resources/templates/`:
- `template1.json` - For API responses
- `template2.json` - For configuration files

# Check if skill is loaded
# In Claude Code:
/plugin list

# Verify skill metadata is correct
# Check SKILL.md frontmatter syntax

# Ensure triggers match user intent
# Add more natural trigger phrases

# If multiple skills trigger, be specific
# Use unique, descriptive names
name: project-specific-testing  # ✅ Good
name: testing                   # ❌ Too generic

# Keep metadata small (<100 tokens in frontmatter)
# Progressive disclosure: Don't load everything at once
# Split large skills into focused sub-skills