Commands

Commands are custom slash commands that appear in your AI IDE. Create reusable prompts with optional arguments to streamline your workflow and ensure consistent AI interactions.

Overview

Commands let you define reusable prompts that can be triggered using slash notation (e.g., /review, /security-check). When you invoke a command, its content is sent to the AI along with any arguments you provide.

Key features include:

• Slash notation — Trigger commands with /command-name in your IDE
• Arguments — Define up to 3 optional or required arguments per command
• Reusable prompts — Write once, use everywhere across your workspace
• Import/Export — Share commands as JSON files

Creating a Command

To create a new command:

1. Navigate to your workspace
2. Click the Commands tab
3. Click New Command
4. Fill in the command details:
   • Name — The command name in kebab-case (e.g., security-review)
   • Description — A brief description of what the command does
   • Content — The prompt text sent to the AI when the command is invoked
5. Click Create

Command Arguments

Commands can have up to 3 arguments that users provide when invoking the command. Arguments make your commands more flexible and reusable.

Adding Arguments

1. Open an existing command or create a new one
2. Click Arguments in the command card menu
3. Click Add Argument
4. Configure the argument:
   • Name — Argument name in kebab-case (e.g., file-path)
   • Description — Help text shown to users
   • Required — Whether the argument must be provided
5. Save the argument

Argument Limits

Using Commands

Once created, commands appear in your IDE when connected via MCP. The exact invocation depends on your IDE:

In Cursor / Claude Code / Windsurf

Use the MCP prompt tool to invoke commands by name. The AI will receive the command content along with any arguments provided.

Enabling & Disabling Commands

Commands can be enabled or disabled:

1. Open the command card menu (three dots)
2. Click Enable or Disable
3. Disabled commands appear dimmed and won't be available in your IDE

Exporting Commands

Commands can be exported as JSON files for backup or sharing:

1. Open the command card menu (three dots)
2. Click Export as JSON
3. The command with all its arguments will be downloaded as a JSON file

The exported JSON format:

{
  "name": "security-review",
  "description": "Perform a security review of the code",
  "content": "Review the following code for security vulnerabilities...",
  "arguments": [
    {
      "name": "focus-area",
      "description": "Specific area to focus on",
      "required": false
    }
  ]
}

Importing Commands

Import commands from JSON files:

1. Click the Import button next to "New Command"
2. Drag and drop a JSON file or click to browse
3. The system validates the JSON structure
4. Preview shows command name, description, and argument count
5. Click Import Command to create it

Note: If a command with the same name exists, it will be auto-renamed (e.g., "review" → "review-2").

Community Commands

Don't want to start from scratch? Browse ready-to-use commands created by the community:

Contributing Your Commands

Have commands that work well for your team? Share them with the community:

1. Export your command as JSON
2. Fork the community repository
3. Add your command to commands/examples/
4. Submit a pull request

See the Contributing Guide for detailed instructions.

Content Limits

Commands have the following content limits:

Workspace Limits

Each workspace can have up to 50 commands. If you reach this limit, you'll need to delete unused commands before creating new ones.