Every Claude Code Command You Need to Know

The first time you type claude in your terminal, the sheer number of available commands can feel overwhelming. There are slash commands for session control, model switching, context management, code review, voice input, and even ordering stickers. Where do you even start?

This guide organizes everything into logical groups so you can learn progressively — starting with the handful of commands you’ll use every single day, then working out to the more specialized power features. Whether you’re a solo developer or part of an enterprise engineering team, this is the reference you’ll want bookmarked.

What Is Claude Code?

Claude Code is Anthropic’s agentic coding tool — not a chat interface with a syntax highlighter, but a full autonomous agent that reads your codebase, edits files, runs terminal commands, manages git workflows, and connects to external services. As of early 2026, roughly 4% of all public GitHub commits are authored by Claude Code, a figure that has grown at a remarkable pace since the tool’s research preview.

Claude Code is available across multiple environments: the terminal CLI, VS Code and JetBrains IDE extensions, a standalone desktop app, and a browser-based interface at claude.ai/code. Most surfaces require a Claude subscription or Anthropic Console account.

The commands in this guide apply primarily to the terminal CLI, which remains the most fully-featured surface. Many slash commands are also accessible in the IDE extensions and desktop app.

Getting Started: Installation and Launch

Claude Code can be installed in three ways. The native binary via the install script is the recommended approach:

# Option 1 — Native binary (recommended, macOS/Linux/WSL)
curl -fsSL https://claude.ai/install.sh | bash

# Option 2 — Homebrew (macOS only)
brew install --cask claude-code

# Option 3 — npm (any OS, legacy — migrate with `claude install`)
npm install -g @anthropic-ai/claude-code

Once installed, launching Claude Code is straightforward. Always start it from inside your project directory:

claude                          # Interactive mode — opens a session in your repo
claude "Fix the bug in auth"    # Start with an initial prompt
claude -p "Explain this module" # One-time query (no persistent session)
claude -c                       # Continue the most recent session

Before your first session, authenticate your account:

claude auth login    # Sign in or switch accounts
claude auth status   # Check your current auth state
claude auth logout   # Clear stored credentials

Session Management Commands

These are the commands you’ll reach for most often during any working session. They control how you navigate, manage, and save your conversations with Claude.

Model and Performance Commands

Claude Code gives you direct control over which model powers your session, and at what effort level. The current model lineup includes Opus 4.7 for complex reasoning, Sonnet 4.6 for general work, and Haiku 4.5 for fast, token-efficient tasks.

A practical decision rule: use Haiku for simple exploration and quick lookups, Sonnet for everyday coding and cost-sensitive work, and Opus for architecture decisions, complex refactoring, agentic loops, and security analysis.

Context and Memory Commands

Managing context is one of the most important skills for working efficiently with Claude Code. These commands help you visualize, optimize, and persist what Claude knows about your project.

Code and Git Commands

These commands are specifically designed for code review, diffing, and integrating Claude Code into your version control workflow.

Configuration and Settings Commands

Claude Code is highly customizable. These commands expose the configuration surface — from visual theming to permissions management to keybinding customization.

Collaboration and Integration Commands

These commands connect Claude Code to your broader toolchain — teammates, mobile devices, MCP servers, and external services.

Bundled Skills

Skills are different from slash commands. While slash commands perform a specific action immediately, skills give Claude a detailed playbook and let it orchestrate complex, multi-step work autonomously — including spawning parallel subagents and managing file output.

Five skills ship with every Claude Code installation:

Keyboard Shortcuts

Knowing these keyboard shortcuts will save you significant time, especially in interactive sessions where you’re rapidly switching modes:

All keybindings are customizable. Run /keybindings to open the configuration file at ~/.claude/keybindings.json. First match wins; Enterprise settings cannot be overridden.

Creating Your Own Commands

One of Claude Code’s most powerful features is the ability to define your own reusable slash commands. These live in .claude/commands/ (project-level) or ~/.claude/commands/ (global) as Markdown files, and they show up automatically in /help.

Commands support YAML frontmatter for descriptions, model overrides, tool restrictions, and argument interpolation:

---
description: Fix a GitHub issue using our coding standards
allowed-tools: Read, Edit, Bash(git:*)
model: opus
argument-hint: [issue-number]
---
Fix GitHub issue #$ARGUMENTS following our conventions.
Check CLAUDE.md for style rules before making any edits.

You can reference files, directories, and even URLs directly in prompts using @ mentions:

@README.md                # Include a specific file
@src/components/          # Include an entire directory
@https://example.com      # Fetch a URL (read-only)

Teams can package frequently-used workflows — like /review-pr or /deploy-staging — as shared skills or commands checked into version control, making them available to every developer on the team.

Putting It All Together

Claude Code’s command surface is large, but you don’t need to master all 60+ commands on day one. Start with the handful that matter most for everyday work — /clear, /compact, /model, /diff, and /init — and let the rest reveal themselves as your workflows grow more sophisticated.

The real leverage comes from combining these primitives: a well-maintained CLAUDE.md feeding Claude project context, hooks enforcing code quality automatically, custom commands packaging team workflows, and MCP servers connecting Claude to your existing tools. That stack turns Claude Code from a smart autocomplete into a genuine development partner.

Claude Code ships weekly, so the command set keeps expanding. Run /release-notes regularly to catch new additions — several of the most useful commands in this guide (like /security-review, /voice, and /color) were added in the past few months alone.