# Slash commands¶
¶
> Control Claude's behavior during an interactive session with slash commands.¶
¶
## Built-in slash commands¶
¶
| Command | Purpose |¶
| :------------------------ | :-------------------------------------------------------------------------------------------------------------------------- |¶
| `/add-dir` | Add additional working directories |¶
| `/agents` | Manage custom AI subagents for specialized tasks |¶
| `/bashes` | List and manage background tasks |¶
| `/bug` | Report bugs (sends conversation to Anthropic) |¶
| `/clear` | Clear conversation history |¶
| `/compact [instructions]` | Compact conversation with optional focus instructions |¶
| `/config` | Open the Settings interface (Config tab) |¶
| `/context` | Visualize current context usage as a colored grid |¶
| `/cost` | Show token usage statistics. See [cost tracking guide](/en/costs#using-the-cost-command) for subscription-specific details. |¶
| `/doctor` | Checks the health of your Claude Code installation |¶
| `/exit` | Exit the REPL |¶
| `/export [filename]` | Export the current conversation to a file or clipboard |¶
| `/help` | Get usage help |¶
| `/hooks` | Manage hook configurations for tool events |¶
| `/ide` | Manage IDE integrations and show status |¶
| `/init` | Initialize project with `CLAUDE.md` guide |¶
| `/install-github-app` | Set up Claude GitHub Actions for a repository |¶
| `/login` | Switch Anthropic accounts |¶
| `/logout` | Sign out from your Anthropic account |¶
| `/mcp` | Manage MCP server connections and OAuth authentication |¶
| `/memory` | Edit `CLAUDE.md` memory files |¶
| `/model` | Select or change the AI model |¶
| `/output-style [style]` | Set the output style directly or from a selection menu |¶
| `/permissions` | View or update [permissions](/en/iam#configuring-permissions) |¶
| `/plan` | Enter plan mode directly from the prompt |¶
| `/plugin` | Manage Claude Code plugins |¶
| `/pr-comments` | View pull request comments |¶
| `/privacy-settings` | View and update your privacy settings |¶
| `/release-notes` | View release notes |¶
| `/rename <name>` | Rename the current session for easier identification |¶
| `/remote-env` | Configure remote session environment (claude.ai subscribers) |¶
| `/resume [session]` | Resume a conversation by ID or name, or open the session picker |¶
| `/review` | Request code review |¶
| `/rewind` | Rewind the conversation and/or code |¶
| `/sandbox` | Enable sandboxed bash tool with filesystem and network isolation for safer, more autonomous execution |¶
| `/security-review` | Complete a security review of pending changes on the current branch |¶
| `/stats` | Visualize daily usage, session history, streaks, and model preferences |¶
| `/status` | Open the Settings interface (Status tab) showing version, model, account, and connectivity |¶
| `/statusline` | Set up Claude Code's status line UI |¶
| `/teleport` | Resume a remote session from claude.ai by session ID, or open a picker (claude.ai subscribers) |¶
| `/terminal-setup` | Install Shift+Enter key binding for newlines (VS Code, Alacritty, Zed, Warp) |¶
| `/theme` | Change the color theme |¶
| `/todos` | List current TODO items |¶
| `/usage` | For subscription plans only: show plan usage limits and rate limit status |¶
| `/vim` | Enter vim mode for alternating insert and command modes |¶
¶
## Custom slash commands¶
¶
Custom slash commands allow you to define frequently used prompts as Markdown files that Claude Code can execute. Commands are organized by scope (project-specific or personal) and support namespacing through directory structures.¶
¶
<Tip>¶
Slash command autocomplete works anywhere in your input, not just at the beginning. Type `/` at any position to see available commands.¶
</Tip>¶
¶
### Syntax¶
¶
```¶
/<command-name> [arguments]¶
```¶
¶
#### Parameters¶
¶
| Parameter | Description |¶
| :--------------- | :---------------------------------------------------------------- |¶
| `<command-name>` | Name derived from the Markdown filename (without `.md` extension) |¶
| `[arguments]` | Optional arguments passed to the command |¶
¶
### Command types¶
¶
#### Project commands¶
¶
Commands stored in your repository and shared with your team. When listed in `/help`, these commands show "(project)" after their description.¶
¶
**Location**: `.claude/commands/`¶
¶
The following example creates the `/optimize` command:¶
¶
```bash theme={null}¶
# Create a project command¶
mkdir -p .claude/commands¶
echo "Analyze this code for performance issues and suggest optimizations:" > .claude/commands/optimize.md¶
```¶
¶
#### Personal commands¶
¶
Commands available across all your projects. When listed in `/help`, these commands show "(user)" after their description.¶
¶
**Location**: `~/.claude/commands/`¶
¶
The following example creates the `/security-review` command:¶
¶
```bash theme={null}¶
# Create a personal command¶
mkdir -p ~/.claude/commands¶
echo "Review this code for security vulnerabilities:" > ~/.claude/commands/security-review.md¶
```¶
¶
### Features¶
¶
#### Namespacing¶
¶
Use subdirectories to group related commands. Subdirectories appear in the command description but don't affect the command name.¶
¶
For example:¶
¶
* `.claude/commands/frontend/component.md` creates `/component` with description "(project:frontend)"¶
* `~/.claude/commands/component.md` creates `/component` with description "(user)"¶
¶
If a project command and user command share the same name, the project command takes precedence and the user command is silently ignored. For example, if both `.claude/commands/deploy.md` and `~/.claude/commands/deploy.md` exist, `/deploy` runs the project version.¶
¶
Commands in different subdirectories can share names since the subdirectory appears in the description to distinguish them. For example, `.claude/commands/frontend/test.md` and `.claude/commands/backend/test.md` both create `/test`, but show as "(project:frontend)" and "(project:backend)" respectively.¶
¶
#### Arguments¶
¶
Pass dynamic values to commands using argument placeholders:¶
¶
##### All arguments with `$ARGUMENTS`¶
¶
The `$ARGUMENTS` placeholder captures all arguments passed to the command:¶
¶
```bash theme={null}¶
# Command definition¶
echo 'Fix issue #$ARGUMENTS following our coding standards' > .claude/commands/fix-issue.md¶
¶
# Usage¶
> /fix-issue 123 high-priority¶
# $ARGUMENTS becomes: "123 high-priority"¶
```¶
¶
##### Individual arguments with `$1`, `$2`, etc.¶
¶
Access specific arguments individually using positional parameters (similar to shell scripts):¶
¶
```bash theme={null}¶
# Command definition ¶
echo 'Review PR #$1 with priority $2 and assign to $3' > .claude/commands/review-pr.md¶
¶
# Usage¶
> /review-pr 456 high alice¶
# $1 becomes "456", $2 becomes "high", $3 becomes "alice"¶
```¶
¶
Use positional arguments when you need to:¶
¶
* Access arguments individually in different parts of your command¶
* Provide defaults for missing arguments¶
* Build more structured commands with specific parameter roles¶
¶
#### Bash command execution¶
¶
Execute bash commands before the slash command runs using the `!` prefix. The output is included in the command context. You *must* include `allowed-tools` with the `Bash` tool, but you can choose the specific bash commands to allow.¶
¶
For example:¶
¶
```markdown theme={null}¶
---¶
allowed-tools: Bash(git add:*), Bash(git status:*), Bash(git commit:*)¶
description: Create a git commit¶
---¶
¶
## Context¶
¶
- Current git status: !`git status`¶
- Current git diff (staged and unstaged changes): !`git diff HEAD`¶
- Current branch: !`git branch --show-current`¶
- Recent commits: !`git log --oneline -10`¶
¶
## Your task¶
¶
Based on the above changes, create a single git commit.¶
```¶
¶
#### File references¶
¶
Include file contents in commands using the `@` prefix to [reference files](/en/common-workflows#reference-files-and-directories).¶
¶
For example:¶
¶
```markdown theme={null}¶
# Reference a specific file¶
¶
Review the implementation in @src/utils/helpers.js¶
¶
# Reference multiple files¶
¶
Compare @src/old-version.js with @src/new-version.js¶
```¶
¶
#### Thinking mode¶
¶
Slash commands can trigger extended thinking by including [extended thinking keywords](/en/common-workflows#use-extended-thinking).¶
¶
### Frontmatter¶
¶
Command files support frontmatter, useful for specifying metadata about the command:¶
¶
| Frontmatter | Purpose | Default |¶
| :------------------------- | :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | :---------------------------------- |¶
| `allowed-tools` | List of tools the command can use | Inherits from the conversation |¶
| `argument-hint` | The arguments expected for the slash command. Example: `argument-hint: add [tagId] \| remove [tagId] \| list`. This hint is shown to the user when auto-completing the slash command. | None |¶
| `context` | Set to `fork` to run the command in a forked sub-agent context with its own conversation history. | Inline (no fork) |¶
| `agent` | Specify which [agent type](/en/sub-agents#built-in-subagents) to use when `context: fork` is set. Only applicable when combined with `context: fork`. | `general-purpose` |¶
| `description` | Brief description of the command | Uses the first line from the prompt |¶
| `model` | Specific model string (see [Models overview](https://docs.claude.com/en/docs/about-claude/models/overview)) | Inherits from the conversation |¶
| `disable-model-invocation` | Whether to prevent the `Skill` tool from calling this command | false |¶
| `hooks` | Define hooks scoped to this command's execution. See [Define hooks for commands](#define-hooks-for-commands). | None |¶
¶
For example:¶
¶
```markdown theme={null}¶
---¶
allowed-tools: Bash(git add:*), Bash(git status:*), Bash(git commit:*)¶
argument-hint: [message]¶
description: Create a git commit¶
model: claude-3-5-haiku-20241022¶
---¶
¶
Create a git commit with message: $ARGUMENTS¶
```¶
¶
Example using positional arguments:¶
¶
```markdown theme={null}¶
---¶
argument-hint: [pr-number] [priority] [assignee]¶
description: Review pull request¶
---¶
¶
Review PR #$1 with priority $2 and assign to $3.¶
Focus on security, performance, and code style.¶
```¶
¶
#### Define hooks for commands¶
¶
Slash commands can define hooks that run during the command's execution. Use the `hooks` field to specify `PreToolUse`, `PostToolUse`, or `Stop` handlers:¶
¶
```markdown theme={null}¶
---¶
description: Deploy to staging with validation¶
hooks:¶
PreToolUse:¶
- matcher: "Bash"¶
hooks:¶
- type: command¶
command: "./scripts/validate-deploy.sh"¶
once: true¶
---¶
¶
Deploy the current branch to staging environment.¶
```¶
¶
The `once: true` option runs the hook only once per session. After the first successful execution, the hook is removed.¶
¶
Hooks defined in a command are scoped to that command's execution and are automatically cleaned up when the command finishes.¶
¶
See [Hooks](/en/hooks) for the complete hook configuration format.¶
¶
## Plugin commands¶
¶
[Plugins](/en/plugins) can provide custom slash commands that integrate seamlessly with Claude Code. Plugin commands work exactly like user-defined commands but are distributed through [plugin marketplaces](/en/plugin-marketplaces).¶
¶
### How plugin commands work¶
¶
Plugin commands are:¶
¶
* **Namespaced**: Commands can use the format `/plugin-name:command-name` to avoid conflicts (plugin prefix is optional unless there are name collisions)¶
* **Automatically available**: Once a plugin is installed and enabled, its commands appear in `/help`¶
* **Fully integrated**: Support all command features (arguments, frontmatter, bash execution, file references)¶
¶
### Plugin command structure¶
¶
**Location**: `commands/` directory in plugin root¶
¶
**File format**: Markdown files with frontmatter¶
¶
**Basic command structure**:¶
¶
```markdown theme={null}¶
---¶
description: Brief description of what the command does¶
---¶
¶
# Command Name¶
¶
Detailed instructions for Claude on how to execute this command.¶
Include specific guidance on parameters, expected outcomes, and any special considerations.¶
```¶
¶
**Advanced command features**:¶
¶
* **Arguments**: Use placeholders like `{arg1}` in command descriptions¶
* **Subdirectories**: Organize commands in subdirectories for namespacing¶
* **Bash integration**: Commands can execute shell scripts and programs¶
* **File references**: Commands can reference and modify project files¶
¶
### Invocation patterns¶
¶
```shell Direct command (when no conflicts) theme={null}¶
/command-name¶
```¶
¶
```shell Plugin-prefixed (when needed for disambiguation) theme={null}¶
/plugin-name:command-name¶
```¶
¶
```shell With arguments (if command supports them) theme={null}¶
/command-name arg1 arg2¶
```¶
¶
## MCP slash commands¶
¶
MCP servers can expose prompts as slash commands that become available in Claude Code. These commands are dynamically discovered from connected MCP servers.¶
¶
### Command format¶
¶
MCP commands follow the pattern:¶
¶
```¶
/mcp__<server-name>__<prompt-name> [arguments]¶
```¶
¶
### Features¶
¶
#### Dynamic discovery¶
¶
MCP commands are automatically available when:¶
¶
* An MCP server is connected and active¶
* The server exposes prompts through the MCP protocol¶
* The prompts are successfully retrieved during connection¶
¶
#### Arguments¶
¶
MCP prompts can accept arguments defined by the server:¶
¶
```¶
# Without arguments¶
> /mcp__github__list_prs¶
¶
# With arguments¶
> /mcp__github__pr_review 456¶
> /mcp__jira__create_issue "Bug title" high¶
```¶
¶
#### Naming conventions¶
¶
Server and prompt names are normalized:¶
¶
* Spaces and special characters become underscores¶
* Names are lowercase for consistency¶
¶
### Managing MCP connections¶
¶
Use the `/mcp` command to:¶
¶
* View all configured MCP servers¶
* Check connection status¶
* Authenticate with OAuth-enabled servers¶
* Clear authentication tokens¶
* View available tools and prompts from each server¶
¶
### MCP permissions and wildcards¶
¶
To approve all tools from an MCP server, use either the server name alone or wildcard syntax:¶
¶
* `mcp__github` (approves all GitHub tools)¶
* `mcp__github__*` (wildcard syntax, also approves all GitHub tools)¶
¶
To approve specific tools, list each one explicitly:¶
¶
* `mcp__github__get_issue`¶
* `mcp__github__list_issues`¶
¶
See [MCP permission rules](/en/iam#tool-specific-permission-rules) for more details.¶
¶
## `Skill` tool¶
¶
<Note>¶
In earlier versions of Claude Code, slash command invocation was provided by a separate `SlashCommand` tool. This has been merged into the `Skill` tool. If you have existing permission rules using `SlashCommand`, update them to use `Skill`.¶
</Note>¶
¶
The `Skill` tool allows Claude to programmatically invoke both [custom slash commands](/en/slash-commands#custom-slash-commands) and [Agent Skills](/en/skills) during a conversation. This gives Claude the ability to use these capabilities on your behalf when appropriate.¶
¶
### What the `Skill` tool can invoke¶
¶
The `Skill` tool provides access to:¶
¶
| Type | Location | Requirements |¶
| :-------------------- | :------------------------------------------- | :--------------------------------------------- |¶
| Custom slash commands | `.claude/commands/` or `~/.claude/commands/` | Must have `description` frontmatter |¶
| Agent Skills | `.claude/skills/` or `~/.claude/skills/` | Must not have `disable-model-invocation: true` |¶
¶
Built-in commands like `/compact` and `/init` are *not* available through this tool.¶
¶
### Encourage Claude to use specific commands¶
¶
To encourage Claude to use the `Skill` tool, reference the command by name, including the slash, in your prompts or `CLAUDE.md` file:¶
¶
```¶
> Run /write-unit-test when you are about to start writing tests.¶
```¶
¶
This tool puts each available command's metadata into context up to the character budget limit. Use `/context` to monitor token usage.¶
¶
To see which commands and Skills are available to the `Skill` tool, run `claude --debug` and trigger a query.¶
¶
### Disable the `Skill` tool¶
¶
To prevent Claude from programmatically invoking any commands or Skills:¶
¶
```bash theme={null}¶
/permissions¶
# Add to deny rules: Skill¶
```¶
¶
This removes the `Skill` tool and all command/Skill descriptions from context.¶
¶
### Disable specific commands or Skills¶
¶
To prevent a specific command or Skill from being invoked programmatically via the `Skill` tool, add `disable-model-invocation: true` to its frontmatter. This also removes the item's metadata from context.¶
¶
<Note>¶
The `user-invocable` field in Skills only controls menu visibility, not `Skill` tool access. Use `disable-model-invocation: true` to block programmatic invocation. See [Control Skill visibility](/en/skills#control-skill-visibility) for details.¶
</Note>¶
¶
### `Skill` permission rules¶
¶
The permission rules support:¶
¶
* **Exact match**: `Skill(/commit)` (allows only `/commit` with no arguments)¶
* **Prefix match**: `Skill(/review-pr:*)` (allows `/review-pr` with any arguments)¶
¶
### Character budget limit¶
¶
The `Skill` tool includes a character budget to limit context usage. This prevents token overflow when many commands and Skills are available.¶
¶
The budget includes each item's name, arguments, and description.¶
¶
* **Default limit**: 15,000 characters¶
* **Custom limit**: Set via `SLASH_COMMAND_TOOL_CHAR_BUDGET` environment variable. The name is retained for backwards compatibility.¶
¶
When the budget is exceeded, Claude sees only a subset of available items. In `/context`, a warning shows how many are included.¶
¶
## Skills vs slash commands¶
¶
**Slash commands** and **Agent Skills** serve different purposes in Claude Code:¶
¶
### Use slash commands for¶
¶
**Quick, frequently used prompts**:¶
¶
* Simple prompt snippets you use often¶
* Quick reminders or templates¶
* Frequently used instructions that fit in one file¶
¶
**Examples**:¶
¶
* `/review` → "Review this code for bugs and suggest improvements"¶
* `/explain` → "Explain this code in simple terms"¶
* `/optimize` → "Analyze this code for performance issues"¶
¶
### Use Skills for¶
¶
**Comprehensive capabilities with structure**:¶
¶
* Complex workflows with multiple steps¶
* Capabilities requiring scripts or utilities¶
* Knowledge organized across multiple files¶
* Team workflows you want to standardize¶
¶
**Examples**:¶
¶
* PDF processing Skill with form-filling scripts and validation¶
* Data analysis Skill with reference docs for different data types¶
* Documentation Skill with style guides and templates¶
¶
### Key differences¶
¶
| Aspect | Slash Commands | Agent Skills |¶
| -------------- | -------------------------------- | ----------------------------------- |¶
| **Complexity** | Simple prompts | Complex capabilities |¶
| **Structure** | Single .md file | Directory with SKILL.md + resources |¶
| **Discovery** | Explicit invocation (`/command`) | Automatic (based on context) |¶
| **Files** | One file only | Multiple files, scripts, templates |¶
| **Scope** | Project or personal | Project or personal |¶
| **Sharing** | Via git | Via git |¶
¶
### Example comparison¶
¶
**As a slash command**:¶
¶
```markdown theme={null}¶
# .claude/commands/review.md¶
Review this code for:¶
- Security vulnerabilities¶
- Performance issues¶
- Code style violations¶
```¶
¶
Usage: `/review` (manual invocation)¶
¶
**As a Skill**:¶
¶
```¶
.claude/skills/code-review/¶
├── SKILL.md (overview and workflows)¶
├── SECURITY.md (security checklist)¶
├── PERFORMANCE.md (performance patterns)¶
├── STYLE.md (style guide reference)¶
└── scripts/¶
└── run-linters.sh¶
```¶
¶
Usage: "Can you review this code?" (automatic discovery)¶
¶
The Skill provides richer context, validation scripts, and organized reference material.¶
¶
### When to use each¶
¶
**Use slash commands**:¶
¶
* You invoke the same prompt repeatedly¶
* The prompt fits in a single file¶
* You want explicit control over when it runs¶
¶
**Use Skills**:¶
¶
* Claude should discover the capability automatically¶
* Multiple files or scripts are needed¶
* Complex workflows with validation steps¶
* Team needs standardized, detailed guidance¶
¶
Both slash commands and Skills can coexist. Use the approach that fits your needs.¶
¶
Learn more about [Agent Skills](/en/skills).¶
¶
## See also¶
¶
* [Plugins](/en/plugins) - Extend Claude Code with custom commands through plugins¶
* [Identity and Access Management](/en/iam) - Complete guide to permissions, including MCP tool permissions¶
* [Interactive mode](/en/interactive-mode) - Shortcuts, input modes, and interactive features¶
* [CLI reference](/en/cli-reference) - Command-line flags and options¶
* [Settings](/en/settings) - Configuration options¶
* [Memory management](/en/memory) - Managing Claude's memory across sessions¶
¶
¶
---¶
¶
> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt