plugin-architecture

Select Claude Code extension types and design plugin.json manifests.

10|2|Updated Mar 12, 2026
One-click install
npx skills add https://github.com/viktorbezdek/skillstack --skill plugin-architecture-viktorbezdek
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: plugin-architecture
Source: https://github.com/viktorbezdek/skillstack/tree/main/plugin-dev/skills/plugin-architecture
Command: npx skills add https://github.com/viktorbezdek/skillstack --skill plugin-architecture-viktorbezdek

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This skill guides developers in deciding which Claude Code extension type—skill, hook, MCP server, subagent, or slash command—to use for a specific capability and helps craft the appropriate plugin.json manifest and directory layout.

Core Features & Use Cases

  • Decision matrix that maps capability triggers to the correct component type.
  • Manifest reference with naming rules, required fields, and multi‑skill layout.
  • Directory layout guidance for organizing skills, hooks, agents, commands, and MCP servers.
  • Real‑world examples drawn from existing plugins to illustrate correct decomposition.

Quick Start

Ask the plugin‑architecture skill to walk you through choosing the right component for your new plugin.

Frequently Asked Questions about plugin-architecture

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I choose the right Claude Code extension type for my plugin capability?

To choose the right Claude Code extension type, map your capability triggers against a decision matrix to select between a skill, hook, MCP server, subagent, or slash command based on your specific requirements.

What is the correct directory layout for organizing Claude Code plugin components?

The correct directory layout for Claude Code plugins organizes skills, hooks, agents, commands, and MCP servers into structured folders to ensure manifest schema adherence and clean component separation.

How do I structure a plugin.json manifest for a multi-skill Claude Code extension?

Structuring a plugin.json manifest for a multi-skill extension requires following specific naming rules and including required fields to define the directory layout and component configurations correctly.

Does this plugin architecture guidance support building both MCP servers and slash commands?

Yes, the plugin architecture guidance supports building MCP servers and slash commands by providing a decision matrix that distinguishes component types and real-world examples for correct decomposition.

What's the best way to decide between using a hook or a subagent in Claude Code plugin development?

The best way to decide between a hook or a subagent is to evaluate your capability triggers using the provided decision matrix, which maps specific execution contexts to the appropriate component type.