Claude Plugin Marketplace

A comprehensive plugin ecosystem for Claude AI, providing extensibility through slash commands, hooks, skills, agents, tools, and Model Context Protocols (MCPs).

🚀 Overview

This repository serves as a starting point for building a Claude plugin marketplace. It provides a well-structured framework for developing, distributing, and managing Claude code plugins with comprehensive .claude and .cursor file systems.

📁 Directory Structure

Plug-in-play/
├── .claude/                    # Claude-specific configurations
│   ├── config.json            # Main configuration
│   ├── schemas/               # JSON schemas for validation
│   │   ├── claude-config.schema.json
│   │   └── plugin-manifest.schema.json
│   └── plugins/               # Plugin directory
│       ├── slash-commands/    # Slash command plugins
│       ├── hooks/             # Lifecycle hook plugins
│       ├── skills/            # Skill-based plugins
│       ├── agents/            # Autonomous agent plugins
│       ├── tools/             # Tool integration plugins
│       └── mcps/              # Model Context Protocol plugins
├── .cursor/                   # Cursor IDE integration
│   ├── settings.json         # Editor and IDE settings
│   └── extensions.json       # Recommended extensions
└── docs/                      # Documentation

🔌 Plugin Categories

1. Slash Commands

Quick-access commands for common operations.

Example: /search, /find-file, /grep

Location: .claude/plugins/slash-commands/

2. Hooks

Lifecycle hooks for context processing and event handling.

Example: pre-process, post-process, on-error

Location: .claude/plugins/hooks/

3. Skills

Advanced capabilities for code analysis, refactoring, and more.

Example: Code quality analysis, security scanning, complexity metrics

Location: .claude/plugins/skills/

4. Agents

Autonomous agents for orchestrating complex multi-step tasks.

Example: Task orchestration, workflow automation, deployment pipelines

Location: .claude/plugins/agents/

5. Tools

Integration with external tools and services.

Example: Git operations, build tools, testing frameworks

Location: .claude/plugins/tools/

6. MCPs (Model Context Protocols)

Enhanced context provision for better AI understanding.

Example: File context, codebase analysis, semantic search

Location: .claude/plugins/mcps/

🛠️ Getting Started

Prerequisites

• Node.js 18+ (for JavaScript plugins)
• Claude AI access
• Cursor IDE (optional, for enhanced development experience)

Installation

1. Clone the repository:

git clone https://github.com/markus41/Plug-in-play.git
cd Plug-in-play

2. Review the configuration:

cat .claude/config.json

3. Explore example plugins:

ls -la .claude/plugins/

📝 Creating a Plugin

Step 1: Choose a Category

Decide which category your plugin belongs to:

• slash-commands - For quick commands
• hooks - For lifecycle events
• skills - For advanced capabilities
• agents - For autonomous operations
• tools - For external integrations
• mcps - For context enhancement

Step 2: Create Plugin Structure

mkdir -p .claude/plugins/[category]/[plugin-name]
cd .claude/plugins/[category]/[plugin-name]

Step 3: Create Manifest

Create a manifest.json file following the schema:

{
  "$schema": "../../schemas/plugin-manifest.schema.json",
  "name": "my-plugin",
  "version": "1.0.0",
  "displayName": "My Plugin",
  "description": "Brief description",
  "category": "tools",
  "author": {
    "name": "Your Name",
    "email": "your.email@example.com"
  },
  "license": "MIT",
  "main": "index.js",
  "entrypoint": "MyPlugin",
  "keywords": ["keyword1", "keyword2"],
  "permissions": ["read", "write"],
  "commands": [],
  "configuration": {}
}

Step 4: Implement Plugin Logic

Create an index.js file with your plugin implementation:

class MyPlugin {
  constructor(context) {
    this.context = context;
    this.config = context.getConfiguration('my-plugin');
  }

  // Implement your plugin logic here
  async execute() {
    // Plugin implementation
  }
}

module.exports = MyPlugin;

Step 5: Add Documentation

Create a README.md file documenting your plugin's usage and configuration.

🔧 Configuration

Claude Configuration (.claude/config.json)

Main configuration for the Claude plugin system:

• marketplace: Marketplace metadata
• plugins: Plugin directory and autoload settings
• settings: Security and validation settings
• api: API endpoint configuration

Cursor Settings (.cursor/settings.json)

IDE integration settings:

• editor: Editor preferences
• claude: Claude integration settings
• pluginDevelopment: Development mode settings
• marketplace: Marketplace UI settings
• security: Security preferences

📚 Example Plugins

Quick Search (Slash Commands)

Fast codebase search capabilities.

Commands:

• /search <pattern> - Search for code patterns
• /find-file <filename> - Find files by name
• /grep <regex> [path] - Advanced regex search

Context Hooks (Hooks)

Lifecycle hooks for enhanced processing.

Hooks:

• pre-process - Before context processing
• post-process - After context processing
• on-error - Error handling
• on-complete - Completion handling

Code Analyzer (Skills)

Advanced code analysis capabilities.

Features:

• Quality metrics
• Complexity analysis
• Security scanning
• Improvement suggestions

Task Orchestrator (Agents)

Intelligent multi-step task automation.

Features:

• Sequential and parallel execution
• Retry logic
• Progress tracking
• Error handling

Git Helper (Tools)

Advanced Git operations.

Features:

• Status checking
• Conventional commits
• History analysis
• Branch management

File Context MCP (MCPs)

Enhanced file and codebase context.

Features:

• File metadata
• Relationship mapping
• Semantic search
• Codebase statistics

🔒 Security

Plugin Validation

All plugins must pass validation checks:

• Schema validation
• Permission verification
• Code signing (optional)
• Security scanning

Permissions

Plugins must declare required permissions:

• read - Read filesystem access
• write - Write filesystem access
• execute - Execute commands
• network - Network access
• filesystem - Full filesystem access

Sandboxing

Plugins run in sandboxed environments with restricted access based on declared permissions.

🤝 Contributing

We welcome contributions! Please see CONTRIBUTING.md for guidelines.

Development Workflow

1. Fork the repository
2. Create a feature branch
3. Develop your plugin
4. Test thoroughly
5. Submit a pull request

📖 Documentation

• Plugin Development Guide
• API Reference
• Contributing Guide
• Security Policy

🗺️ Roadmap

• Plugin registry API
• Web-based marketplace UI
• Plugin discovery and search
• Automated testing framework
• Plugin signing and verification
• Usage analytics
• Plugin ratings and reviews
• Community-driven plugins

📄 License

MIT License - see LICENSE for details.

🙏 Acknowledgments

Built for the Claude AI community to enhance development capabilities and productivity.

📞 Support

• GitHub Issues: Report bugs or request features
• Discussions: Ask questions and share ideas

---

Happy Plugin Development! 🎉