Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/VineeTagarwal-code/claude-code/llms.txt

Use this file to discover all available pages before exploring further.

Slash commands let you control Claude Code’s behavior, navigate sessions, configure tools, and access information without leaving the terminal. Type / at the prompt to see a filtered list of available commands.
The set of commands available to you depends on your account type and provider. Commands that require a claude.ai subscription or direct Anthropic API access may not appear when using Bedrock, Vertex, or third-party providers.

​
Session & Navigation

Aliases: /reset, /newClears the current conversation history and frees up context, starting a fresh session without restarting the process.
/clear
Use this when you want to start a new task from scratch. For keeping a summary while discarding the history, use /compact instead.
Usage: /compact [instructions]Clears conversation history but preserves a generated summary in context. Reduces token usage while retaining the important context Claude needs to continue working. You can optionally pass custom summarization instructions.
/compact
/compact focus on the authentication changes we made
This command supports non-interactive (pipeline) mode. It can be disabled by setting DISABLE_COMPACT=1.
Aliases: /continueUsage: /resume [conversation id or search term]Opens a picker to resume a previous conversation. Pass a conversation ID or a search term to filter the list.
/resume
/resume auth refactor
Aliases: /remoteDisplays the URL and QR code for the current remote session. Only available when running in remote mode (CLAUDE_CODE_REMOTE=true).
/session
Aliases: /quitExits the Claude Code REPL immediately.
/exit
Usage: /rename [name]Renames the current conversation. If you omit the name, Claude Code prompts you to enter one interactively.
/rename
/rename auth-refactor-session
Usage: /copy [N]Copies Claude’s most recent response to the clipboard. Pass a number N to copy the Nth-latest response instead.
/copy
/copy 2
Aliases: /checkpointRestores the code and/or conversation to a previous point in the session, allowing you to undo recent changes.
/rewind
Usage: /branch [name]Creates a branch of the current conversation at the current point, letting you explore a different direction without affecting the original thread.
/branch
/branch try-different-approach

​
AI & Model

Usage: /model [model]Opens the model picker or directly sets the model Claude uses. Shows the currently active model in the command description.
/model
/model claude-opus-4-5
Usage: /effort [low|medium|high|max|auto]Controls how much thinking effort Claude applies to responses. Higher effort levels use more tokens and take longer but produce more thorough results.
/effort high
/effort auto
LevelBehavior
lowMinimal thinking, fastest responses
mediumBalanced thinking and speed
highExtended thinking for complex problems
maxMaximum thinking budget
autoClaude decides based on task complexity
Usage: /plan [open|description]Enables plan mode or views the current session plan. In plan mode, Claude proposes a plan before taking actions, letting you review and approve before execution.
/plan
/plan open
Usage: /fast [on|off]Toggles fast mode, which uses a smaller, faster model for responses. Only available for claude.ai subscribers and direct Anthropic API users when fast mode is enabled for your account.
/fast on
/fast off

​
Tools & Permissions

Aliases: /allowed-toolsOpens the permissions management panel where you can view, add, and remove allow/deny rules for tools. Permission rules control which shell commands and file operations Claude can run without asking.
/permissions
Permission rules use patterns like Bash(git *) to allow all git commands, or Edit(src/**) to allow edits within the src/ directory.
Opens the hooks viewer to inspect hook configurations for tool events. Hooks let you run custom scripts before or after tool calls.
/hooks
Hook configurations live in your settings.json or project-level CLAUDE.md files. See /help for documentation links.
Usage: /mcp [enable|disable [server-name]]Opens the MCP server management panel. You can enable, disable, and inspect connected MCP (Model Context Protocol) servers from here.
/mcp
/mcp enable filesystem
/mcp disable puppeteer
Usage: /ide [open]Shows the status of IDE integrations (VS Code, JetBrains, etc.) and lets you manage connections. Pass open to launch the IDE integration setup.
/ide
/ide open

​
Content & Context

Opens an editor for Claude’s memory files (e.g., CLAUDE.md). Memory files provide persistent context that Claude reads at the start of every session.
/memory
Renders a visual grid showing how much of the available context window is currently in use. Useful for understanding when you are approaching the context limit and should run /compact.
/context
In non-interactive mode, outputs a plain-text summary of context usage instead.
Usage: /add-dir <path>Adds a new working directory to Claude’s context. Claude can then read and modify files in that directory.
/add-dir /path/to/project
/add-dir ../sibling-repo
Shows uncommitted changes in the current repository and displays per-turn diffs of what Claude has modified during the session.
/diff

​
Skills & Agents

Lists all skills available in the current session. Skills are reusable prompt-based commands that can be loaded from the ~/.claude/skills/ directory, plugin directories, or bundled with Claude Code.
/skills
Opens the agent configuration manager where you can view and edit agent definitions. Agents are specialized sub-processes Claude can launch to handle specific tasks.
/agents
Aliases: /plugins, /marketplaceOpens the plugin marketplace and management panel. You can browse, install, enable, disable, and configure Claude Code plugins.
/plugin
Aliases: /bashesLists all currently running and recent background tasks (shell commands and agents launched in the background). You can view output and manage tasks from here.
/tasks

​
Information & Diagnostics

Displays the help screen with available commands and keyboard shortcuts.
/help
Shows a comprehensive status panel including the current version, active model, account information, API connectivity, and tool statuses.
/status
Runs a series of diagnostic checks to verify your Claude Code installation and configuration. Reports installation type, version, path issues, auto-update status, multiple-installation conflicts, and context warnings (large CLAUDE.md files, MCP tool token counts, unreachable permission rules).
/doctor
The doctor command checks:
  • Installation type and path (npm global, npm local, native, package manager)
  • Whether ~/.local/bin is in PATH for native installs
  • Auto-update eligibility and permissions
  • Multiple conflicting installations
  • Large CLAUDE.md files consuming context
  • MCP tool schema token usage
  • Unreachable permission rules
  • ripgrep availability and mode
Displays the total API cost and duration of the current session. This command is hidden for claude.ai subscribers who do not pay per token.
/cost
Shows your Claude Code usage statistics and activity over time.
/stats
Shows your current plan usage and rate limits. Only available for claude.ai subscribers.
/usage

​
Appearance

Opens the theme picker so you can change the color theme of the Claude Code terminal UI.
/theme
Usage: /color <color|default>Sets the color of the prompt bar for the current session. Pass a named color or hex value, or default to reset.
/color blue
/color #ff6b6b
/color default
Toggles between Vim and Normal editing modes for the input field.
/vim
Opens or creates your keybindings configuration file. Only available when keybinding customization is enabled.
/keybindings
Usage: /statusline [description]Launches an agent to configure Claude Code’s status line UI, optionally guided by a description of what you want. If no argument is given, it reads your shell PS1 configuration to set up a matching status line.
/statusline
/statusline show git branch and last command exit code
Aliases: /settingsOpens the configuration panel where you can adjust Claude Code settings.
/config

​
Workflows

Triggers a code review of the current changes or a specified pull request.
/review
Aliases: /bugUsage: /feedback [report]Opens a feedback dialog to submit product feedback or bug reports to Anthropic. Not available for Bedrock, Vertex, or Foundry API users.
/feedback
/feedback report
Prompts you to upgrade to a Max plan for higher rate limits and more Opus model access. Only available for claude.ai subscribers who are not on an enterprise plan.
/upgrade