Skip to main content

Anthropic Claude Code: Official Implementation & User Guidance Resources

Complete reference of Anthropic's official documentation, guides, and resources for Claude Code

🎯 Core Documentation Hub​

Official Documentation Home​

https://code.claude.com/docs/en/overview

  • Main documentation portal
  • Product overview and capabilities
  • Installation methods and requirements

Quickstart Guide (5 minutes)​

https://code.claude.com/docs/en/quickstart

  • Fast-track installation and setup
  • First H.P.002-COMMANDS to try
  • Understanding your codebase
  • Making first code changes

📚 Getting Started Resources​

Installation & Setup​

https://code.claude.com/docs/en/setup

  • Multiple installation methods (native, Homebrew, npm)
  • Platform-specific instructions (macOS, Linux, Windows)
  • Authentication setup
  • Troubleshooting installation issues

Common Workflows​

https://code.claude.com/docs/en/tutorials

  • Understanding new codebases
  • Bug fixing efficiently
  • Code refactoring
  • Working with tests
  • Creating pull requests
  • Documentation handling
  • Extended thinking usage
  • Git worktrees for parallel sessions

🏗️ Architecture & Implementation​

Best Practices for Agentic Coding​

https://www.anthropic.com/engineering/claude-code-best-practices

  • Official Anthropic engineering blog post
  • Environment tuning and optimization
  • CLAUDE.md usage patterns
  • MCP integration strategies
  • Custom command H.P.008-TEMPLATES
  • Token efficiency techniques
  • Multi-step H.P.006-WORKFLOWS

How Anthropic Teams Use Claude Code​

https://www-cdn.anthropic.com/58284b19e702b49db9302d5b6f135ad8871e7658.pdf

  • Internal team case studies
  • Real-world usage patterns
  • Data pipeline navigation
  • Codebase discovery H.P.006-WORKFLOWS
  • Performance optimizations

🛠️ Configuration & Settings​

Settings Documentation​

https://code.claude.com/docs/en/settings

  • Configuration hierarchy and scopes
  • settings.json reference
  • Environment variables
  • Permission system
  • Tool H.P.009-CONFIGuration
  • Hooks system overview

CLAUDE.md Memory System​

  • Project context and instructions
  • Coding standards documentation
  • Architecture guidelines
  • File organization patterns
  • Custom H.P.006-WORKFLOWS

🧩 Extensibility Features​

Agent Skills​

https://code.claude.com/docs/en/H.P.003-SKILLS

  • What Skills are and when to use them
  • Creating SKILL.md files
  • Skill frontmatter H.P.009-CONFIGuration
  • Supporting files and H.P.004-SCRIPTS
  • Tool restrictions
  • Dynamic context injection
  • Distribution options

Deep Dive: Equipping Agents with Agent Skills (Referenced in documentation - comprehensive architectural guide)

Custom SubH.P.001-AGENTS​

https://code.claude.com/docs/en/sub-H.P.001-AGENTS

  • Built-in subH.P.001-AGENTS (Plan, Research, Explore)
  • Creating custom subH.P.001-AGENTS
  • Model selection for subH.P.001-AGENTS
  • Tool and permission control
  • Foreground vs background execution
  • Delegation patterns

Hooks System​

https://code.claude.com/docs/en/H.P.005-HOOKS-guide

  • Hook events overview (PreToolUse, PostToolUse, Stop)
  • Creating H.P.005-HOOKS for automation
  • Code formatting H.P.005-HOOKS
  • Custom notifications
  • File protection patterns
  • Security checks

Hooks Reference https://code.claude.com/docs/en/H.P.005-HOOKS

  • Complete hook H.P.009-CONFIGuration reference
  • Event types and triggers
  • Matcher patterns
  • MCP tool integration with H.P.005-HOOKS
  • Component-scoped H.P.005-HOOKS

Slash Commands​

https://code.claude.com/docs/en/slash-H.P.002-COMMANDS (Note: URL refers to Skills but includes slash command info)

  • Creating custom H.P.002-COMMANDS
  • Command frontmatter
  • Argument passing
  • User-invocable vs model-invocable

🔌 Model Context Protocol (MCP)​

MCP Integration Guide​

https://code.claude.com/docs/en/mcp

  • What MCP enables
  • Popular MCP servers (Sentry, GitHub, Figma, databases)
  • Installing MCP servers (HTTP, SSE, stdio)
  • Scope management (local, project, user)
  • Dynamic tool updates
  • Tool Search optimization
  • Output token management

Custom Tools (MCP in SDK)​

https://code.claude.com/docs/en/sdk/custom-tools

  • Creating custom MCP servers
  • Type-safe tool definitions
  • Tool naming conventions
  • Selective tool allowlisting
  • Streaming input requirements

📦 Plugin System​

Discover and Install Plugins​

https://code.claude.com/docs/en/discover-plugins

  • Official plugin marketplace
  • Installing prebuilt plugins
  • Plugin management H.P.002-COMMANDS
  • User vs project plugins

Create Plugins​

https://code.claude.com/docs/en/plugins

  • Plugin architecture overview
  • Creating plugin manifest (plugin.json)
  • Plugin directory structure
  • Adding H.P.002-COMMANDS, Skills, H.P.005-HOOKS, MCP servers
  • Testing plugins locally
  • Debugging plugin issues
  • Converting existing H.P.009-CONFIGs to plugins

Plugins Reference​

https://code.claude.com/docs/en/plugins-reference

  • Complete technical specifications
  • Plugin.json schema
  • Directory structure requirements
  • Slash H.P.002-COMMANDS in plugins
  • Agents in plugins
  • Skills in plugins
  • Hooks H.P.009-CONFIGuration
  • MCP server bundling
  • LSP server integration

Official Plugin Repository​

https://github.com/anthropics/claude-plugins-official

  • Anthropic-managed plugins
  • Example plugin implementations
  • Plugin submission process
  • Quality and security standards

🚀 Deployment & Integration​

VS Code Extension​

https://code.claude.com/docs/en/vs-code

  • Installing Claude Code in VS Code
  • Graphical interface features
  • @-mention functionality
  • Plan review and editing
  • Auto-accept mode
  • Keyboard shortcuts
  • Third-party provider H.P.009-CONFIGuration (Bedrock, Vertex AI)

GitHub Actions​

https://code.claude.com/docs/en/github-actions

  • Setting up GitHub Actions integration
  • Tagging @claude in issues/PRs
  • Workflow H.P.009-CONFIGuration
  • Breaking changes from beta to v1.0
  • Direct vs tag mode

GitLab CI/CD​

https://code.claude.com/docs/en/gitlab-cicd (Referenced in navigation - CI/CD integration)

Claude Code on the Web​

https://code.claude.com/docs/en/claude-code-on-the-web

  • Web-based Claude Code sessions
  • Session handoff (terminal ↔ web)
  • Background task execution
  • Cloud environment H.P.009-CONFIGuration
  • Monitoring from mobile app
  • & prefix for web sessions
  • Teleporting sessions

Claude Code in Slack​

https://code.claude.com/docs/en/slack (Referenced in navigation - Slack integration)

📖 CLI & Reference​

CLI Reference​

https://code.claude.com/docs/en/cli-reference

  • Command-line arguments
  • Interactive mode features
  • Non-interactive usage
  • Session management

Interactive Mode​

https://code.claude.com/docs/en/interactive-mode

  • Keyboard shortcuts (macOS/Linux/Windows)
  • Text editing shortcuts
  • Vim mode
  • Multiline input (Shift+Enter)
  • Background bash H.P.002-COMMANDS
  • Command history (Ctrl+R)

Slash Commands Reference​

  • Complete list of built-in H.P.002-COMMANDS
  • /help, /status, /clear, /exit
  • /allowed-tools, /permissions
  • /H.P.001-AGENTS, /H.P.005-HOOKS, /plugin
  • Project and user command scopes

Programmatic Usage (SDK)​

https://code.claude.com/docs/en/sdk/

  • Using Claude Code programmatically
  • SDK documentation
  • Automation and scripting

🎓 Learning Resources​

Anthropic Academy​

https://www.anthropic.com/learn/build-with-claude

  • Claude API development guide
  • Quickstart repositories
  • Cookbook code snippets
  • Full API documentation

Changelog​

https://code.claude.com/docs/en/changelog

  • Version history
  • New features by release
  • Breaking changes
  • Migration guides

🐛 Troubleshooting & Support​

Troubleshooting Documentation​

(Referenced throughout docs - issue-specific solutions)

GitHub Repository​

https://github.com/anthropics/claude-code

  • Official repository
  • Bug reporting (/bug command or GitHub issues)
  • Feature requests
  • Community discussions

Support Channels​

🎨 Customization Features​

Output Styles​

https://code.claude.com/docs/en/output-styles (Referenced in navigation)

  • Customizing Claude's output formatting
  • Style H.P.008-TEMPLATES
  • Project-specific styles

Extended Thinking​

  • Deep reasoning for complex problems
  • Triggering modes: "think", "think hard", "think harder", "ultrathink"
  • Viewing thinking process
  • When to use extended thinking
  • Tab key toggle during sessions

Permission Modes​

  • Default mode (ask for permission)
  • Plan mode (read-only, creates plan before acting)
  • Auto-accept mode (accepts all changes)
  • Configuring default mode
  • Switching modes mid-session

📂 File & Directory Management​

Configuration Hierarchy​

Precedence (High to Low):

  1. Managed Settings (Enterprise)

    • macOS: /Library/Application Support/ClaudeCode/managed-settings.json
    • Linux: /etc/claude-code/managed-settings.json
    • Windows: C:\Program Files\ClaudeCode\managed-settings.json

  2. Command-line Arguments

  3. Local Project Settings

    • .claude/settings.local.json (not committed to git)

  4. Shared Project Settings

    • .claude/settings.json (committed to git)

  5. User Settings

    • ~/.claude/settings.json (applies to all projects)

  6. Built-in Defaults

File Locations Reference​

File TypeUser (Global)Project (Shared)Project (Local)
Settings~/.claude/settings.json.claude/settings.json.claude/settings.local.json
Memory~/.claude/CLAUDE.mdCLAUDE.mdCLAUDE.local.md
Commands~/.claude/H.P.002-COMMANDS/.claude/H.P.002-COMMANDS/-
Agents~/.claude/H.P.001-AGENTS/.claude/H.P.001-AGENTS/-
Skills~/.claude/H.P.003-SKILLS/.claude/H.P.003-SKILLS/-
MCP~/.claude.json (merged).mcp.json-

🔐 Security & Compliance​

Permission System​

  • Tool-based permissions (Read, Write, Bash, WebFetch)
  • Allowlist and denylist patterns
  • Directory restrictions
  • Command restrictions
  • WebFetch domain controls

Attribution Configuration​

  • Git commit attribution
  • Pull request attribution
  • Customizing attribution messages
  • Disabling attribution

Sandbox Mode​

  • Isolated execution environment
  • Network restrictions
  • Command exclusions
  • Unix socket allowlisting

🌟 Advanced Features​

Multi-Project Management​

  • Git worktrees for parallel sessions
  • Running multiple Claude instances
  • Project isolation strategies

Unix-Style Utility Usage​

  • Piping input/output
  • Non-interactive mode
  • Scripting with Claude Code
  • Integration with shell pipelines

Checkpoint & Resume​

  • Session checkpointing
  • --continue flag
  • --resume for specific conversations
  • Session history management

Image Handling​

  • Viewing images inline
  • Image analysis capabilities
  • Screenshot integration

🔢 Model Selection​

Available Models​

Opus 4.5 - Most capable

  • Model string: claude-opus-4-5-20251101
  • Use for: Complex reasoning, architectural decisions

Sonnet 4.5 - Balanced (default)

  • Model string: claude-sonnet-4-5-20250929
  • Use for: General coding, most tasks

Haiku 4.5 - Fastest

  • Model string: claude-haiku-4-5-20251001
  • Use for: Simple tasks, quick responses

Model Configuration​

  • Set default in settings.json
  • Override per project
  • Specify for subH.P.001-AGENTS
  • Switch during session with /model

💡 Key Concepts Summary​

Core Principles​

Agentic Coding Claude Code acts autonomously, making plans, executing changes, and verifying results with minimal user intervention.

Context Awareness Automatically reads CLAUDE.md files, indexes project structure, and maintains awareness of your entire codebase.

Permission-First Always asks permission before file modifications or command execution (unless in auto-accept mode).

Extensibility Skills, subH.P.001-AGENTS, H.P.005-HOOKS, MCP servers, and plugins allow deep customization.

Token Efficiency Progressive disclosure, lazy loading, Tool Search, and context management optimize token usage.

🎯 Implementation Philosophy​

Anthropic's Approach​

1. Natural Language Interface Describe tasks in plain English rather than learning complex H.P.002-COMMANDS.

2. Plan → Execute → Verify Three-phase workflow for complex changes ensures quality.

3. Progressive Disclosure Information provided incrementally as needed, not all upfront.

4. Tool Composability MCP servers, Skills, and H.P.005-HOOKS work together seamlessly.

5. Team-First Design Project-level H.P.009-CONFIGuration enables team-wide standards.

📊 Usage Patterns​

Workflow Archetypes​

Research & Understanding

claude chat
> Analyze this codebase and explain the authentication flow

Feature Development

claude chat
> Add pagination to the user list API endpoint

Bug Fixing

claude chat
> Fix the race condition in UserService.updateProfile()

Refactoring

claude chat
> Refactor auth module to use async/await consistently

Documentation

claude chat
> Update README with new deployment instructions

🔗 Essential Links Quick Reference​

ResourceURL
Main Docshttps://code.claude.com/docs/en/overview
Quickstarthttps://code.claude.com/docs/en/quickstart
Best Practiceshttps://www.anthropic.com/engineering/claude-code-best-practices
Settingshttps://code.claude.com/docs/en/settings
Skillshttps://code.claude.com/docs/en/H.P.003-SKILLS
MCPhttps://code.claude.com/docs/en/mcp
Pluginshttps://code.claude.com/docs/en/plugins
GitHub Repohttps://github.com/anthropics/claude-code
VS Codehttps://code.claude.com/docs/en/vs-code
GitHub Actionshttps://code.claude.com/docs/en/github-actions

📝 Documentation Notes​

Documentation Characteristics​

Comprehensive Coverage Every feature has dedicated documentation with examples and best practices.

Progressive Complexity Starts simple (quickstart) → intermediate (common H.P.006-WORKFLOWS) → advanced (SDK, plugins).

Multi-Format Learning

  • Step-by-step guides
  • Reference documentation
  • Code examples
  • Video tutorials (planned)
  • Interactive examples

Living Documentation Changelog tracks all updates, new features documented immediately.

Getting Started
├── Overview
├── Setup
├── Quickstart
└── Common Workflows

Build with Claude Code
├── SubH.P.001-AGENTS
├── Plugins
├── Skills
├── Hooks
├── MCP
└── SDK

Deployment
├── VS Code
├── JetBrains
├── GitHub Actions
├── GitLab CI/CD
└── Slack Integration

Administration
├── Settings
├── Managed Settings
└── Enterprise Deployment

Configuration
└── (Settings, Permissions, Tools)

Reference
├── CLI Reference
├── Interactive Mode
├── Slash Commands
├── Hooks Reference
└── Plugins Reference

Resources
└── Changelog

🎓 Learning Path Recommendation​

Beginner (Day 1)​

  1. Read Overview
  2. Follow Quickstart
  3. Try Common Workflows
  4. Understand Permission System

Intermediate (Week 1)​

  1. Configure settings.json
  2. Create CLAUDE.md for your project
  3. Install 2-3 MCP servers
  4. Create first custom command

Advanced (Month 1)​

  1. Create custom Skills
  2. Build custom subagent
  3. Set up H.P.005-HOOKS for automation
  4. Create shareable plugin

Expert (Ongoing)​

  1. Contribute to official plugins
  2. Build MCP servers
  3. Implement CI/CD integration
  4. Develop team-wide standards

🚀 Next Steps​

After reviewing this guide:

  1. Install Claude Code

    curl -fsSL https://claude.ai/install.sh | bash

  2. Complete Quickstart https://code.claude.com/docs/en/quickstart

  3. Read Best Practices https://www.anthropic.com/engineering/claude-code-best-practices

  4. Set Up Your Project

    • Create CLAUDE.md
    • Configure permissions
    • Install relevant MCP servers

  5. Join Community

    • GitHub Discussions
    • Follow Changelog
    • Share your Skills/Plugins

📚 Related Resources​

Model Context Protocol​

https://modelcontextprotocol.io/

  • Official MCP specification
  • Building MCP servers
  • MCP server registry

Anthropic API Documentation​

https://docs.anthropic.com/

  • Claude API reference
  • Prompt engineering
  • Best practices

Anthropic Console​

https://console.anthropic.com/

  • API key management
  • Usage tracking
  • Workspace management

Last Updated: January 2026
Documentation Version: Claude Code 2.x
Maintained By: Anthropic

Note: This guide aggregates links and information from Anthropic's official Claude Code documentation. For the most current information, always refer to https://code.claude.com/docs/