development-agents

Personal Claude Code Plugin Marketplace

A minimal template for creating and organizing your own Claude Code plugins.

Overview

This repository provides a basic framework for developing Claude Code plugins. It includes a simple example plugin that demonstrates the core plugin architecture.

What's Included

• Example Plugin - A minimal plugin with basic examples of:
  • Agent (hello-agent)
  • Command (/hello)
  • Skill (code_style)
  • MCP server configuration

Quick Start

1. Clone and Customize

# Clone or copy this repository
git clone <your-repo-url>
cd marketplace

# Update marketplace configuration
# Edit .claude-plugin/marketplace.json with your name

2. Create Your First Plugin

# Copy the example plugin as a template
cp -r example my-plugin

# Update plugin metadata
# Edit my-plugin/plugin.json and my-plugin/.claude-plugin/plugin.json

3. Register Your Plugin

Add your plugin to .claude-plugin/marketplace.json:

{
  "plugins": [
    {
      "name": "my-plugin",
      "source": "./my-plugin",
      "description": "Your plugin description"
    }
  ]
}

4. Test in Claude Code

# Start Claude Code in this directory
claude code

# Test your command
/hello

Plugin Structure

marketplace/
├── .claude-plugin/
│   └── marketplace.json          # Plugin registry
├── example/                       # Example plugin
│   ├── .claude-plugin/
│   │   └── plugin.json           # Plugin metadata
│   ├── agents/
│   │   └── hello-agent.md        # Agent definition
│   ├── commands/
│   │   └── hello.md              # Command definition
│   ├── skills/
│   │   └── code_style.md         # Skill definition
│   ├── .mcp.json                 # MCP server config (optional)
│   └── plugin.json               # Plugin metadata
└── README.md

Plugin Components

Commands

Commands are user-facing actions invoked via /command-name.

Structure:

---
description: Brief description of what this command does
---

# Command Instructions

Detailed instructions for Claude on how to respond...

File location: plugin/commands/command-name.md

Agents

Agents are specialized autonomous workers for complex tasks.

Structure:

---
name: agent-name
description: When to use this agent
model: sonnet
color: blue
---

# Agent Prompt

Core competencies and instructions...

File location: plugin/agents/agent-name.md

Skills

Skills are reusable patterns and expertise that Claude can reference.

Structure:

---
name: Skill Name
description: When to use this skill
---

# Skill Content

Guidelines, patterns, and examples...

File location: plugin/skills/skill-name.md

MCP Servers (Optional)

MCP (Model Context Protocol) servers enable integration with external services.

Structure:

{
  "mcpServers": {
    "ServerName": {
      "command": "npx|docker",
      "args": ["..."],
      "env": {
        "ENV_VAR": "${SYSTEM_ENV_VAR}"
      }
    }
  }
}

File location: plugin/.mcp.json

Development Workflow

Adding a Command

1. Create plugin/commands/my-command.md
2. Add YAML frontmatter with description
3. Write instructions for Claude
4. Test with /my-command

Adding an Agent

1. Create plugin/agents/my-agent.md
2. Add YAML frontmatter (name, description, model, color)
3. Write agent prompt with core competencies
4. Agent is automatically available

Adding a Skill

1. Create plugin/skills/my-skill.md
2. Add YAML frontmatter with name and description
3. Write skill content (patterns, guidelines, examples)
4. Reference in commands or agent prompts

Adding MCP Integration

1. Create or update plugin/.mcp.json
2. Add server configuration with command and args
3. Use environment variables for secrets
4. Restart Claude Code session to load MCP servers

Best Practices

Naming Conventions

• Commands: lowercase-with-hyphens.md → /command-name
• Agents: descriptive-name.md
• Skills: snake_case.md
• Plugins: lowercase-with-hyphens

Security

• Never commit credentials - use environment variables
• Reference env vars: ${VAR_NAME} in configuration
• Use minimal required permissions for service accounts

Organization

• Keep plugins focused on a specific domain or workflow
• Group related commands, agents, and skills together
• Use clear, descriptive names for all components

Example Plugin

The included example plugin demonstrates:

• hello-agent - Shows basic agent structure with YAML frontmatter
• /hello command - Demonstrates simple command with instructions
• code_style skill - Shows how to define reusable guidelines
• .mcp.json - Template for MCP server configuration

Use these as templates for your own plugins.

Troubleshooting