Claude Code Guide
The definitive zero-to-hero guide for Claude Code.
Who This Guide Is For
Whether you're a seasoned developer, a vibe coder building with AI, or someone who's never touched a terminal before, this guide meets you where you are. No programming experience required.
Chapters
Introduction & Getting Started
Learn how Claude Code's agentic loop works, install on macOS/Linux/Windows, explore the core tool set, and complete your first AI-assisted coding session.
Project Context & Memory
Master the CLAUDE.md hierarchy, auto-memory system, path-specific rules, and context window management to get dramatically better results from Claude Code.
Models, Cost Economics & Permissions
Choose between Sonnet, Opus, and Haiku models with 1M extended context. Configure effort levels with /effort, track spending, and set up the 6-mode permission system including Auto Mode.
Environment Sandboxing & Customization
Configure Claude Code's four-tier settings hierarchy with managed-settings.d/ policy fragments, enable sandbox isolation, and customize the terminal display with --bare mode and flicker-free rendering.
Remote Control & Headless Automation
Run Claude Code from any device via remote control with multi-session server mode, interact through Telegram, Discord, and iMessage channels, automate with headless mode in CI/CD, and schedule recurring tasks.
Model Context Protocol
Connect Claude Code to external tools — databases, APIs, ticketing systems — using the Model Context Protocol with stdio, HTTP, and OAuth authentication.
Custom Skills
Build reusable SKILL.md files with conditional path-based activation, content lifecycle management across compaction, and distribution through the Plugins system.
Hooks & Lifecycle Automation
Automate quality enforcement with 24 lifecycle events, conditional if-field filtering, the PermissionDenied event for auto mode integration, and defer decisions for SDK workflows.
Git Worktrees & Subagent Delegation
Parallelize development across isolated git worktree branches using built-in and custom subagents with AGENT.md personas, persistent memory, initialPrompt auto-submit, and sparse-checkout for monorepos.
Agent Teams & Advanced Orchestration
Coordinate multiple Claude Code instances on shared task lists with dependency tracking and peer-to-peer communication (research preview).
Security & Enterprise Administration
Deploy Claude Code at organizational scale with Auto Mode governance, 6 permission modes, managed-settings.d/ policies, sandbox enforcement, Bash hardening, and audit compliance.
Plugins
Build shareable plugin packages with manifests, bin/ executables, userConfig, hooks, and marketplace publishing for distribution across teams and the community.
Agent SDK
Use the Python and TypeScript SDKs to embed Claude Code's agent loop, tools, and context management into production applications and CI/CD pipelines.
Computer Use
Control native desktop applications through Claude Code with per-app approval, safety-first design, and CLI or Desktop GUI setup on macOS and Windows.