Plugin Categories
Han plugins are organized into four categories: Core, Jutsu, Do, and Hashi. Each serves a specific purpose in the quality enforcement ecosystem.
Han plugins follow Japanese martial arts naming conventions, organizing by purpose. Each category serves a specific role in the quality enforcement ecosystem.
Core - Foundation
The essential infrastructure that powers Han. Always required.
What's included:
- Quality enforcement through validation hooks
- Metrics tracking and confidence calibration
- Context7 integration for up-to-date library documentation
- Universal programming principles (SOLID, DRY, composition over inheritance)
- MCP servers for hooks and metrics
- Binary auto-installation on first session
Key plugins:
- core (
han-core) - The technical foundation. Provides hooks, metrics, and all core infrastructure - bushido - Optional philosophical layer based on seven Samurai virtues (義 Righteousness, 勇 Courage, 仁 Compassion, 礼 Respect, 誠 Honesty, 名誉 Honor, 忠義 Loyalty)
When to install:
Always. Core is required for Han to function. Bushido is optional—install only if its philosophical approach resonates with you. All technical capabilities come from core.
Installation:
han plugin install core
# Optional philosophy layer
han plugin install bushido
Jutsu (術) - Technical Skills
Jutsu plugins are "techniques"—deep knowledge of specific technologies paired with automatic validation.
What's included:
- Technology-specific expertise (skills, commands)
- Validation hooks that run automatically
- Best practices and patterns
- Error detection and remediation guidance
Examples:
- jutsu-typescript - TypeScript expertise + type checking hooks
- jutsu-playwright - E2E testing knowledge + test validation
- jutsu-nextjs - Next.js patterns + build verification
- jutsu-biome - Code formatting + automatic linting
- jutsu-python - Python skills + validation
- jutsu-rust - Rust patterns + compilation checks
When to install:
Install jutsu plugins for every technology in your stack. They ensure Claude not only knows the technology but validates its work automatically.
Installation:
# Auto-detect and install all relevant jutsu plugins
han plugin install --auto
# Or install specific technologies
han plugin install jutsu-typescript
han plugin install jutsu-react
han plugin install jutsu-python
Dō (道) - Specialized Agents
Do plugins provide specialized agents for complex, multi-phase workflows. Think of them as expert consultants with deep domain knowledge.
What's included:
- Autonomous agents for specific disciplines
- Multi-step workflows
- Domain-specific best practices
- Quality checklists and verification
Examples:
- do-frontend-development - UI/UX-focused agent with accessibility expertise
- do-technical-documentation - Documentation agent following best practices
- do-accessibility-engineering - Multiple agents for inclusive design
- do-code-review - Comprehensive code review with confidence-based filtering
- do-debugging - Systematic debugging workflows
- do-architecture-design - System architecture and planning
When to install:
Install do plugins for specialized tasks you perform regularly. Each agent brings deep expertise and handles complexity autonomously.
Installation:
# Browse available agents
han plugin search do-
# Install specific agents
han plugin install do-frontend-development
han plugin install do-code-review
Hashi (橋) - External Bridges
Hashi plugins are MCP servers that connect Claude to external services and tools. They turn Claude into a universal interface for your development workflow.
What's included:
- MCP server implementations
- Authentication and authorization
- API integrations
- Tool-specific commands and workflows
Examples:
- hashi-github - GitHub Issues, PRs, code search, Actions
- hashi-playwright-mcp - Browser automation and testing
- hashi-blueprints - Codebase documentation and knowledge management
- hashi-jira - Issue tracking and project management
- hashi-sentry - Error tracking and monitoring
When to install:
Install hashi plugins for external services you use in your workflow. They enable Claude to interact with these services naturally through conversation.
Installation:
# Install to user settings (recommended for MCP servers)
han plugin install hashi-github
han plugin install hashi-playwright-mcp
# Or specify scope explicitly
han plugin install hashi-blueprints --scope user
How They Work Together
Han plugins compose into a complete quality system. Here's a real example:
Request: "Add user authentication to the app"
What happens:
- Core provides infrastructure and quality enforcement
- jutsu-nextjs provides Next.js implementation knowledge
- jutsu-typescript ensures type safety throughout
- do-frontend-development handles UI components
- Validation hooks run automatically (via core):
- TypeScript compilation check
- Next.js build verification
- Test suite execution
- Core code review analyzes the result
- hashi-github can create a PR with changes
All of this happens automatically from one request. No manual intervention needed.
Installation Scopes
Plugins can be installed to different scopes:
- user (default) - Shared across all projects (
~/.claude/settings.json) - project - Team settings for current project (
.claude/settings.json) - local - Personal overrides, gitignored (
.claude/settings.local.json)
Recommendations:
- user scope: MCP servers (hashi-), core plugins, general agents (do-)
- project scope: Technology validation (jutsu-* with hooks)
- local scope: Personal preferences not shared with team
Next Steps
- Install Han and auto-detect your stack
- Browse the plugin marketplace
- Learn about configuration
- Explore the CLI reference