What are Claude Code Plugins?

Claude Code Plugins are shareable packages that bundle skills, hooks, MCP servers, agents, and LSP servers together. They let you extend Claude Code with custom functionality that can be shared across projects and teams.

How are plugins different from standalone skills?

Standalone skills live in your .claude/ directory and are best for personal or project-specific workflows. Plugins are directories with a .claude-plugin/plugin.json manifest, designed for sharing with teammates, community distribution, versioned releases, and reuse across projects.

What can a plugin contain?

A plugin can contain skills (SKILL.md files), commands (slash commands), hooks (event handlers), MCP server configurations, LSP server configurations, custom agents, and default settings.

How do I create a Claude Code plugin?

Create a directory with a .claude-plugin/plugin.json manifest containing name, description, version, and author. Then add skills, hooks, MCP servers, or agents in their respective directories. Test locally with the --plugin-dir flag.

What are Plugins? - Complete Guide to Claude Code Plugins

Plugins are an official feature of Claude Code that let you bundle skills, hooks, MCP servers, agents, and LSP servers into shareable packages. Think of them as "extension packs" that give Claude Code new capabilities you can share across projects and teams.

compare_arrowsPlugins vs Standalone Configuration

Approach	Skill Names	Best For
Standalone `.claude/`	`/hello`	Personal workflows, project-specific, quick experiments
Plugin `.claude-plugin/`	`/plugin-name:hello`	Team sharing, community distribution, versioned releases

inventory_2What a Plugin Can Contain

check_circleSkills — SKILL.md files with specialized instructions and workflows
check_circleCommands — Slash commands as markdown files, namespaced by plugin
check_circleHooks — Event handlers that run on Claude Code lifecycle events
check_circleMCP Servers — Model Context Protocol servers for external tool integrations
check_circleLSP Servers — Language Server Protocol servers for real-time code intelligence
check_circleAgents — Custom agent definitions for specialized workflows
check_circleSettings — Default configuration applied when the plugin is enabled

folder_openPlugin Structure

foldermy-plugin/

my-plugin/
  .claude-plugin/
    plugin.json           # Manifest (name, description, version, author)
  commands/               # Slash commands as Markdown files
  skills/                 # Agent Skills with SKILL.md files
  agents/                 # Custom agent definitions
  hooks/                  # Event handlers (hooks.json)
  .mcp.json               # MCP server configurations
  .lsp.json               # LSP server configurations
  settings.json           # Default settings applied when enabled

descriptionExample plugin.json Manifest

description.claude-plugin/plugin.json

{
  "name": "my-first-plugin",
  "description": "A greeting plugin to learn the basics",
  "version": "1.0.0",
  "author": {
    "name": "Your Name"
  }
}

The manifest identifies your plugin and is required. The name field becomes the namespace prefix for all slash commands and skills within the plugin.

codeExample SKILL.md Inside a Plugin

descriptionskills/hello/SKILL.md

---
description: Greet the user with a friendly message
disable-model-invocation: true
---

Greet the user warmly and ask how you can help them today.

Skills inside a plugin are namespaced. If this lives in a plugin named my-first-plugin, the skill is invoked as /my-first-plugin:hello. The $ARGUMENTS placeholder can capture user input after the command name for dynamic behavior.

build_circleHow to Create a Plugin

Create the plugin manifest

Create a directory with a .claude-plugin/plugin.json file containing your plugin's name, description, version, and author information.

Add skills, hooks, or servers

Populate the plugin with the components you need: add SKILL.md files in skills/, event handlers in hooks/, MCP servers in .mcp.json, or custom agents in agents/.

Test locally

Use the --plugin-dir ./my-plugin flag when launching Claude Code to load your plugin for testing. You can load multiple plugins with repeated --plugin-dir flags.

Share on Mirdan.ai

Once your plugin is working well, share it with the community! Others can install it with /plugin install or through the Mirdan CLI.

add_circleSubmit Your Plugin

swap_horizMigrating from Standalone to Plugin

Already have skills and hooks in your .claude/ directory? You can convert them into a plugin:

Create a new directory with a .claude-plugin/plugin.json manifest
Move your skill files into the plugin's skills/ directory
Move hooks and MCP configs into their respective locations
Test with --plugin-dir to verify everything works

Note: Skill names will be namespaced with your plugin name after migration (e.g., /tdd becomes /my-plugin:tdd).

menu_bookOfficial Anthropic Documentation

Plugins are an official feature of Claude Code by Anthropic (requires v1.0.33 or later). For the most up-to-date information on plugin structure, capabilities, and best practices, refer to the official documentation:

open_in_newClaude Code Documentation open_in_newPlugins Reference Guide

Ready to explore Plugins?

Browse our curated collection of community and official plugins for Claude Code.

exploreBrowse Plugins