Claude Code Slash Commands

Claude Code Slash Commands

Complete reference for all built-in and custom slash commands.

Built-in Slash Commands

Session Management

Command	Description
/clear	Clear conversation and start fresh (same session)
/compact	Compress conversation to reduce context usage
/compact [instructions]	Compact with custom focus instructions
/resume	Resume a previous session
/continue	Continue current session
/status	Current session info
/context	View context window usage

Help & Info

Command	Description
/help	Show available commands
/doctor	Run diagnostics and health checks
/version	Show Claude Code version
/settings	View active settings

Model & Configuration

Command	Description
/model	View or change the current model
/model <name>	Switch to specific model (sonnet, opus, haiku)
/permissions	Manage permission rules

Conversation Modes

Command	Description
/plan	Enter plan mode (read-only analysis)
/fast	Toggle fast mode (faster output, same model)

Memory & Instructions

Command	Description
/memory	Edit CLAUDE.md and auto-memory
/init	Generate project CLAUDE.md
/init-only	Generate CLAUDE.md and exit

Agents & Skills

Command	Description
/agents	View, create, edit, delete subagents
/agents list	List all available agents
/skills	Browse available skills
/plugin	Manage plugins (install, enable, disable)

Tools & Servers

Command	Description
/mcp	Check MCP server status, authenticate
/hooks	Configure lifecycle hooks
/add-dir <path>	Add additional working directory

Code & Navigation

Command	Description
/diff	View git diff interactively
/review	Review pull request
/fork [name]	Fork current conversation

Display & Preferences

Command	Description
/theme	Change color theme
/vim	Toggle Vim editing mode
/terminal-setup	Configure keybindings

Built-in Skills

Command	Description
/simplify	Review code quality and refactor
/batch <instruction>	Parallel codebase changes
/debug [description]	Debug session issues
/loop <interval>	Recurring prompts
/claude-api	Load Claude API reference docs

Security & Sandbox

Command	Description
/sandbox	View or configure sandbox settings
/security-review	Run security analysis on codebase

Export & Sharing

Command	Description
/export	Export conversation
/copy	Copy last response to clipboard

Monitoring & Insights

Command	Description
/tasks	View background tasks
/insights	Session insights and analytics
/stats	Usage statistics
/usage	Token/cost usage for current session
/extra-usage	Extended usage details

Releases & Updates

Command	Description
/release-notes	View release notes for current version
/upgrade	Upgrade Claude Code to latest

Hooks & Plugins

Command	Description
/save-hooks	Save current hooks configuration
/reload-plugins	Reload installed plugins

Browser & Display

Command	Description
/chrome	Open Chrome browser session
/desktop	Desktop view mode
/mobile	Mobile view mode
/ide	IDE integration mode

Account & Auth

Command	Description
/login	Login to Anthropic
/logout	Logout from Anthropic

Configuration

Command	Description
/rename	Rename current conversation
/keybindings	View or edit keybindings
/statusline	Configure status line display
/output-style	Change output formatting style
/privacy-settings	View or edit privacy settings
/add-dir	Add working directory to context

Remote & Apps

Command	Description
/remote-env	Configure remote environment
/remote-control	Start remote control session

GitHub & Slack Apps

Command	Description
/install-github-app	Install Claude GitHub App
/install-slack-app	Install Claude Slack App

Bug Reporting

Command	Description
/bug	Report a bug (can be disabled via env var)

Input Prefixes

Prefix	Action
/	Invoke command or skill
!	Bash mode (execute shell command)
@	File mention (add file to context)
?	Show shortcuts

Keyboard Shortcuts

Input Controls

Shortcut	Action
Enter	Send message
Shift+Enter	Multi-line input
Ctrl+C	Interrupt current operation
Ctrl+D	Exit Claude Code
Escape	Rewind to last checkpoint
Escape (2x)	Full rewind
Tab	Auto-complete
Up/Down	Navigate command history
?	Show shortcuts

Mode & Toggle Controls

Shortcut	Action
Shift+Tab	Cycle permission modes
Ctrl+O	Toggle verbose output
Ctrl+B	Send task to background
Ctrl+T	Toggle task list
Option+P	Switch model
Option+T	Toggle extended thinking
Ctrl+G	Open input in text editor
Ctrl+V	Paste image

Line Editing

Shortcut	Action
Ctrl+K	Kill text from cursor to end of line
Ctrl+U	Kill text from cursor to start of line
Ctrl+Y	Yank (paste) killed text
Alt+Y	Cycle through kill ring
Alt+B	Move back one word
Alt+F	Move forward one word
Alt+M	Toggle multiline mode

Display & Navigation

Shortcut	Action
Ctrl+L	Clear screen
Ctrl+R	Reverse search history
Ctrl+F	Kill background agents

Permission Prompts

Shortcut	Action
y / Enter	Approve (Allow Once)
a	Allow Always (for this session)
n	Deny
d	Deny Always (for this session)
e	Edit (modify the tool call)

Custom Slash Commands

Creating Custom Commands

Custom commands are skill files in .claude/skills/ or plugin directories.

Skill File Format

---
name: my-command
description: What this command does
triggers:
  - /my-command
  - /mc
args:
  - name: target
    description: Target to operate on
    required: false
---

# My Custom Command

Instructions for Claude when this command is invoked.

## Steps
1. Do this first
2. Then do this
3. Finally do this

## Rules
- Always follow this pattern
- Never do this other thing

Plugin Command Format

Commands in plugin commands/ directories:

# Command Name

Description of what the command does.

## Usage
\`\`\`
/plugin-prefix:command-name [args]
\`\`\`

## Arguments
- `arg1` — Description
- `--flag` — Description

## Implementation

When this command is invoked:
1. Step one
2. Step two
3. Step three

Command Resolution Order

1. Built-in commands (highest priority)
2. Project skills (.claude/skills/)
3. Installed plugin commands
4. User-level skills (~/.claude/skills/)

Invoking Custom Commands

# Direct invocation
/my-command

# With arguments
/my-command some-argument

# Plugin-namespaced command
/plugin-name:command-name

# Via Skill tool (programmatic)
Skill(skill="my-command", args="some-argument")

Plan Mode Details

When /plan is active:

1. Claude analyzes the request
2. Writes a step-by-step plan to a plan file
3. Calls ExitPlanMode to present plan to user
4. User reviews and approves/modifies/rejects
5. On approval, Claude executes the plan

Plan Mode Permissions

• Read-only tools work normally
• Write/Edit/Bash require plan approval first
• Use AskUserQuestion for clarifications during planning
• Use ExitPlanMode when plan is ready (not AskUserQuestion)

Context Management Commands

/compact

Compresses the conversation to free up context window space.

/compact                    # Default compression
/compact focus on the API   # Compress with focus on API-related content

When to use:

• Context window is getting large
• Conversation has covered many topics
• You want to focus on a specific area

How it works:

• Summarizes previous conversation
• Preserves key decisions and context
• Reduces token count significantly
• Maintains file state awareness

/clear

Starts a completely fresh conversation.

/clear

When to use:

• Switching to unrelated task
• Want a clean slate
• Context is too polluted

Difference from /compact:

• /compact preserves summarized context
• /clear removes all context entirely