Claude Code SDK

Authoritative reference for Claude Code extensibility and configuration.

Route to Reference

Working with	Reference
Skills: SKILL.md, frontmatter, activation, subagent execution	skills.md
Subagents: custom agents, tool restrictions, delegation	subagents.md
Plugins: plugin.json, directory structure, distribution	plugins.md
Hooks: lifecycle events, command/prompt/agent hooks	hooks.md
MCP: servers, tools, resources, authentication	mcp.md
Memory: CLAUDE.md, rules, imports, hierarchy	memory.md
Output Styles: personas, system prompt modification	output-styles.md
Settings: scopes, permissions, sandbox, env vars, tools	settings.md
Status line: custom display, JSON input, scripts	statusline.md
Usage patterns: context management, verification, scaling	best-practices.md

Read the relevant reference before making changes.

Concepts

<concepts> **Skill** — Prompt template in `SKILL.md` that extends Claude's capabilities. Loaded on-demand when description matches user request. Can include `references/` for detailed content. Invoked with `/skill-name` or automatically.

Plugin — Distributable package containing skills, hooks, MCP servers, output styles. Has .claude-plugin/plugin.json manifest. Installed from marketplace or local path. Skills namespaced as /plugin:skill.

Hook — Deterministic automation triggered at lifecycle events (tool use, session start, permission request). Three types: command (shell), prompt (LLM decision), agent (multi-turn verification). Configured in settings or plugin.

MCP Server — External tool/resource provider via Model Context Protocol. Connects Claude to databases, APIs, services. Configured per-project, per-user, or via plugin.

Output Style — Persona/behavior modifier via system prompt changes. Affects how Claude responds without changing capabilities. Built-in: Default, Explanatory, Learning.

CLAUDE.md — Project memory file providing persistent context about codebase, conventions, instructions. Hierarchy: managed → user → project. Loaded automatically at session start.

Subagent — Isolated context for delegated tasks. Built-in types: Explore (read-only), Plan (architecture), general-purpose. Custom agents in .claude/agents/.

Settings — Configuration hierarchy controlling permissions, model, hooks, behavior. Scopes: managed → user → project → local. </concepts>

Quick Reference

Skill Locations

Scope	Path
Personal	`~/.claude/skills/<name>/SKILL.md`
Project	`.claude/skills/<name>/SKILL.md`
Plugin	`<plugin>/skills/<name>/SKILL.md`

Plugin Layout

code

plugin-name/
├── .claude-plugin/
│   └── plugin.json        # Manifest (required)
├── skills/
│   └── skill-name/
│       ├── SKILL.md
│       └── references/
├── hooks/
│   └── hooks.json
├── output-styles/
├── .mcp.json
└── README.md

Hook Events

Event	When	Can Block
`PreToolUse`	Before tool	Yes
`PostToolUse`	After tool success	No
`Stop`	Claude finishes	Yes
`SessionStart`	Session begins	No

Settings Scopes

Scope	Location	Shared
Managed	System path	Org-wide
User	`~/.claude/settings.json`	No
Project	`.claude/settings.json`	Git
Local	`.claude/settings.local.json`	No

When to Use This Skill

•Creating plugins, skills, hooks, or output styles
•Debugging activation failures or hook issues
•Understanding Claude Code configuration hierarchy
•Setting up MCP servers or tool integration
•Writing or modifying CLAUDE.md files
•Any question about Claude Code internals