Claude Code CLI Commands: Complete Cheat Sheet (2026)

Claude Code CLI Commands: The Complete Cheat Sheet for 2026

Claude Code is Anthropic's official command-line interface for interacting with Claude directly in your terminal and codebase. Whether you are using it for code generation, debugging, refactoring, or full-feature development, knowing the CLI commands and flags is essential for maximizing your productivity.

This cheat sheet covers every major command, slash command, flag, and configuration option available in Claude Code as of early 2026.

Installation

Install Claude Code globally via npm:

npm install -g @anthropic-ai/claude-code

Verify the installation:

claude --version

Update to the latest version:

npm update -g @anthropic-ai/claude-code

Basic Usage

Launch Claude Code in interactive mode from any directory:

claude

Send a one-shot prompt without entering interactive mode:

claude "explain what this project does"

Pipe input into Claude Code:

cat error.log | claude "explain this error and suggest a fix"

Resume the most recent conversation:

claude --resume

CLI Flags Reference

Flag	Short	Description
--version	-v	Print the installed version
--help	-h	Show help information
--resume	-r	Resume the most recent conversation
--continue	-c	Continue the last conversation
--print	-p	Print mode: respond without interactive session
--model	-m	Specify the model to use
--output-format		Set output format: text, json, stream-json
--max-turns		Limit the number of agentic turns
--system-prompt		Provide a custom system prompt
--allowedTools		Restrict available tools
--disallowedTools		Exclude specific tools
--mcp-config		Path to MCP servers config file
--permission-mode		Set permission mode: default, plan, auto-edit
--verbose		Enable verbose logging output
--dangerously-skip-permissions		Skip all permission prompts (use with caution)

Examples of Flag Usage

Run Claude in print mode with a specific model:

claude -p -m claude-sonnet-4 "write a Python function to parse CSV files"

Limit agentic turns to prevent runaway operations:

claude --max-turns 5 "refactor the auth module"

Use JSON output for scripting:

claude -p --output-format json "list all TODO comments in this project"

Run with a custom system prompt:

claude --system-prompt "You are a security auditor. Review code for vulnerabilities." "audit src/"

Slash Commands (Interactive Mode)

Once inside an interactive Claude Code session, you can use slash commands:

Command	Description
/help	Show available commands and usage information
/clear	Clear the conversation history and start fresh
/compact	Condense the conversation to save context window
/config	View or modify configuration settings
/cost	Display token usage and estimated cost for the session
/doctor	Run diagnostics to check for configuration issues
/init	Initialize a CLAUDE.md file in the current project
/login	Authenticate with your Anthropic account
/logout	Sign out of the current session
/mcp	View or manage MCP server connections
/model	Switch the active model mid-conversation
/permissions	View or modify tool permissions
/review	Request a code review of staged changes
/status	Show current session status and configuration
/terminal-setup	Configure terminal integration (keybindings, shell)
/vim	Toggle vim keybinding mode

The CLAUDE.md File

Claude Code reads a CLAUDE.md file from your project root to understand project-specific context. Initialize one with:

claude /init

Or create it manually. A typical CLAUDE.md includes:

# Project Overview
This is a Next.js 14 application using TypeScript and Tailwind CSS.

# Tech Stack
- Framework: Next.js 14 (App Router)
- Language: TypeScript
- Styling: Tailwind CSS + shadcn/ui
- Database: PostgreSQL via Prisma
- Auth: NextAuth.js v5

# Coding Conventions
- Use functional components with hooks
- Prefer named exports
- Use absolute imports with @/ prefix
- Write tests with Vitest

# File Structure
- src/app/ - Next.js app router pages
- src/components/ - Reusable UI components
- src/lib/ - Utility functions and shared logic
- prisma/ - Database schema and migrations

Configuration

View your current configuration:

claude /config

Configuration is stored in ~/.claude/ and includes:

File	Purpose
~/.claude/settings.json	Global user settings
~/.claude/mcp_servers.json	MCP server configurations
.claude/settings.json	Project-level settings (commit to repo)
CLAUDE.md	Project context file
~/.claude/CLAUDE.md	Global context file (applies to all projects)

Permission Modes

Claude Code has three permission modes that control how much autonomy the AI has:

Mode	Behavior
default	Asks permission for file writes and command execution
plan	Read-only mode; Claude can analyze but not modify files
auto-edit	Automatically approves file edits, still asks for commands

Set the mode when launching:

claude --permission-mode plan "analyze the architecture of this project"

Or configure specific tool permissions in .claude/settings.json:

{
  "permissions": {
    "allow": [
      "Read",
      "Glob",
      "Grep",
      "Bash(npm test)",
      "Bash(npm run lint)"
    ],
    "deny": [
      "Bash(rm *)",
      "Bash(git push)"
    ]
  }
}

MCP Server Management

List connected MCP servers:

claude /mcp

Add an MCP server via the config file at ~/.claude/mcp_servers.json:

{
  "github": {
    "command": "npx",
    "args": ["@anthropic/mcp-server-github"],
    "env": {
      "GITHUB_TOKEN": "ghp_xxxxxxxxxxxx"
    }
  }
}

Useful Patterns and Workflows

Quick code review before committing:

git diff --staged | claude -p "review these changes for bugs and suggest improvements"

Generate commit messages:

git diff --staged | claude -p "write a concise commit message for these changes"

Explain an unfamiliar codebase:

claude "give me a high-level overview of this project's architecture"

Run tests and fix failures:

claude "run the test suite and fix any failing tests"

Batch processing with scripts:

for file in src/components/*.tsx; do
  claude -p "add JSDoc comments to all exported functions" < "$file"
done

Keyboard Shortcuts (Interactive Mode)

Shortcut	Action
Ctrl+C	Cancel the current generation
Ctrl+D	Exit Claude Code
Up Arrow	Scroll through prompt history
Tab	Accept autocomplete suggestions
Esc	Toggle between single-line and multi-line input
Shift+Tab	Reject the current suggestion

Cost Monitoring

Track your spending during a session:

/cost

This displays:

• Total input tokens consumed
• Total output tokens generated
• Estimated cost in USD
• Cache hit rate (for repeated context)

Troubleshooting

Claude Code is not reading my CLAUDE.md: Ensure the file is in the root of your working directory. Claude Code looks for CLAUDE.md in the current directory and parent directories.

MCP servers fail to connect: Check that the command specified in your MCP config is available on your PATH. Run the command manually to verify it works.

High token usage: Use /compact regularly to compress conversation history. This reduces context window usage and costs.

Permission denied errors: Check your .claude/settings.json permissions configuration. You may need to add specific tools to the allow list.

Conclusion

Claude Code is one of the most powerful AI coding assistants available in 2026, and mastering its CLI commands gives you full control over your AI-powered development workflow. Keep this cheat sheet bookmarked for quick reference.

If you are building applications that need AI media generation -- images, videos, audio, or talking avatars -- check out Hypereal AI. Hypereal offers a unified API with pay-as-you-go pricing, fast generation times, and access to the latest models like Flux, Kling, and Wan 2.2.