Commands

Create, manage, and organize custom slash commands for Claude Code.

Overview

Slash commands are markdown files that Claude Code executes when you type /command-name. Claude Deck lets you browse all available commands across user, project, and plugin scopes, create new ones with a guided wizard, and edit existing ones.

How to Use

Browsing Commands

The Commands page shows all available commands grouped by scope:

• User — global commands in ~/.claude/commands/
• Project — project-specific commands in .claude/commands/
• Plugin — commands provided by installed plugins (read-only)

Use the search bar to filter by name, description, or content. Stats at the top show counts per scope.

Creating a Command

Click "Add Command" to open the wizard:

1. Name — enter the command name and optional namespace (e.g., tools:analyze becomes /tools:analyze)
2. Scope — choose user (global) or project (version-controlled)
3. Frontmatter — add description and allowed tools
4. Content — select a template or write custom markdown

Editing Commands

Click a command card to view its details, then click Edit to modify the description, allowed tools, or markdown content.

Allowed Tools

Commands can restrict which tools they use via the allowed-tools frontmatter:

---
description: Run the test suite
allowed-tools: Bash, Read
---
Run all tests and report results.

Available tools: Read, Write, Edit, Bash, Glob, Grep, WebFetch, Task

Configuration

Location	Scope	Description
~/.claude/commands/	User	Global commands
.claude/commands/	Project	Project-specific commands

Commands are stored as .md files. Namespaces map to subdirectories: tools:analyze → commands/tools/analyze.md.

Tips

• Plugin commands are read-only — you can view them but not edit or delete.
• Namespaces help organize large command libraries (e.g., git:hooks, tools:lint).
• Use allowed-tools to restrict what a command can do for security.