Where should I store my plugin components?

By convention, commands go in /commands, skills in /skills/{name}/SKILL.md, and agents in /agents, as mapped in your plugin.json file.

What is the difference between a Command and a Skill?

Commands are user-invoked slash commands defined in Markdown, while Skills are automated knowledge sets that Claude applies contextually during a session.

Can I specify an AI model inside a Skill definition?

No, models cannot be specified for Skills. However, they can be specified as full IDs for Commands and short names for Agents.

What is the mandatory file for a Claude Code plugin?

The plugin.json file is required in the root directory and must contain at least the 'name' field in kebab-case.

Claude Code Plugin Architect

Name: Claude Code Plugin Architect
Author: biwakonbu

bybiwakonbu

開発者ツール

Guides the creation and validation of Claude Code plugins by providing comprehensive specification knowledge for commands, skills, and agents.

概要

The Plugin Spec skill serves as a centralized knowledge hub for developers building extensions for the Claude Code CLI. It integrates detailed technical specifications for plugin.json manifests, slash commands, automated skills, sub-agents, and event hooks into a single accessible interface. By providing structural validation rules and implementation patterns, this skill ensures that developers can architect robust, compliant plugins that follow official documentation while avoiding common pitfalls like incorrect model IDs or missing metadata.

主な機能

Unified overview of commands, skills, agents, and hooks architecture
0 GitHub stars
Comprehensive plugin.json schema guidance and mandatory field validation
Component-specific model selection rules and naming conventions
Pre-defined validation rules for plugin directory structures
Delegation to specialized creation skills for rapid scaffolding

ユースケース

Validating existing plugin configurations against the latest official specifications
Architecting a new Claude Code plugin from scratch with correct file mappings
Clarifying the structural differences between skills, agents, and slash commands

概要

主な機能

Unified overview of commands, skills, agents, and hooks architecture
0 GitHub stars
Comprehensive plugin.json schema guidance and mandatory field validation
Component-specific model selection rules and naming conventions
Pre-defined validation rules for plugin directory structures
Delegation to specialized creation skills for rapid scaffolding

ユースケース

Validating existing plugin configurations against the latest official specifications
Architecting a new Claude Code plugin from scratch with correct file mappings
Clarifying the structural differences between skills, agents, and slash commands