Claude Code Plugin Marketplace
A collection of plugins, commands, skills, agents, and hooks for Claude Code - extending Claude's capabilities for development workflows.
Quick Start
Add this marketplace to your Claude Code installation:
/plugin marketplace add shawn-sandy/claude-code
Then install any plugin:
/plugin install starter-plugin@claude-code-marketplace
Local Development
For local testing or development, you can add this marketplace from your filesystem:
# Add marketplace from local directory
/plugin marketplace add /path/to/claude-code
# Or use a relative path
/plugin marketplace add ./path/to/claude-code
# Then install plugins
/plugin install starter-plugin@claude-code-marketplace
This is useful when:
- Testing new plugins before publishing
- Developing plugins locally
- Using private/internal plugins not on GitHub
Git Hooks
This repository uses Husky git hooks for automated validation. The hooks run automatically and require no npm/node setup:
Validation Hooks:
pre-commit - Validates JSON, shell scripts, and YAML frontmatter
commit-msg - Enforces conventional commit format (<type>: <description>)
pre-push - Runs OpenSpec validation and security checks
prepare-commit-msg - Provides helpful commit message template
The hooks use system tools (jq, bash, openspec) without additional dependencies.
Commit Message Format:
# Required format
<type>: <description>
<type>(<scope>): <description>
# Examples
docs: update README with installation steps
feat(plugins): add new testing plugin
fix: resolve JSON validation issue
For troubleshooting and bypassing hooks, see CONTRIBUTING.md.
Available Plugins
starter-plugin
A comprehensive example plugin demonstrating all Claude Code component types with detailed documentation and best practices.
Components:
- Commands: Example slash command (
/example) with frontmatter configuration
- Skills: Autonomous skill that explains skill development
- Agents: Specialized agent for learning about agent structure
- Hooks: Event-driven automation examples (PreToolUse, PostToolUse, SessionStart)
- MCP: Configuration templates for Model Context Protocol servers
Install:
/plugin install starter-plugin@claude-code-marketplace
Usage Examples:
# Try the example command
/example Hello from the marketplace!
# Skills are automatically triggered by Claude based on context
# Try asking: "Show me how to create a skill"
# Agents are launched via the Task tool
# Try asking: "Use the example-agent to explain agents"
plugin-dev
Plugin development toolkit with interactive skills for creating, scaffolding, and managing Claude Code plugins.
Components:
- Skills:
plugin-setup - Interactive plugin scaffolding with comprehensive validation
Install:
/plugin install plugin-dev@claude-code-marketplace
Usage Examples:
# The plugin-setup skill triggers automatically when you say:
"Create a new plugin"
"Setup a plugin"
"Add a skill to my plugin"
"Create a command"
# The skill will guide you through:
# 1. Gathering plugin requirements
# 2. Creating proper directory structure
# 3. Generating component templates
# 4. Updating marketplace catalog
# 5. Running comprehensive validation
# 6. Providing local testing instructions
Plugin Components Explained
Commands (Slash Commands)
User-invoked actions that start with /. Perfect for:
- Running tests or builds
- Creating git commits
- Custom workflows
- Deployment operations
Structure: commands/command-name.md with YAML frontmatter
Skills
Autonomously triggered capabilities that enhance Claude's abilities. Perfect for:
- Code analysis and formatting
- Documentation generation
- Pattern detection
- Specialized knowledge
Structure: skills/skill-name/SKILL.md with YAML frontmatter
Best Practice: Complex skills should include README.md for user documentation separate from the system prompt. See templates/skill-readme-template.md for template.
Agents
Specialized subagents for complex, domain-specific tasks. Perfect for:
- Code reviews
- Security audits
- Research and analysis
- Multi-step workflows
Structure: agents/agent-name.md with YAML frontmatter
Hooks
Event-driven automation that responds to tool usage and session events. Perfect for:
- Validation and safety checks
- Logging and monitoring
- Enforcing best practices
- Custom workflows
Structure: hooks/hooks.json with hook configurations
MCP Servers
Model Context Protocol integrations for external services. Perfect for:
- API integrations
- Database access
- External tool integration
- Custom protocols
Structure: .mcp.json with server configurations
Management Commands
Marketplace Commands