Cook with AI Enrollment closed
AI & LLMs Runs locally

Slash Command Generator

Build custom slash command markdown files for Cursor and Claude Code, with arguments, allowed tools, and the prompt body.

← All tools No signup · Nothing uploaded
ELI5

It creates shortcut commands you can type to give your AI a common task.

A little more detail

What this tool does

Custom slash commands are markdown files stored in your project or home directory. Claude Code supports frontmatter for allowed tools and argument hints; Cursor reads plain command files. This generator builds the file for your target tool and shows where to save it.

Use it to
  • Turn a prompt you keep retyping into a one-word command
  • Pass arguments into a command template with placeholders
  • Limit which tools a command is allowed to use in Claude Code
Interactive workspace Results update as you type
Target

Claude Code uses YAML frontmatter in .claude/commands/. Cursor uses plain markdown in .cursor/commands/.

Invoke as · Save to

One line for YAML frontmatter. Shown in command pickers.

Optional note line under the command title in the markdown file.

Shown in the command UI. Use placeholders like [branch-name] for expected args.

Optional. Omit to inherit conversation permissions. Restrict tools this command may call, comma-separated.

Prompt body

Use $ARGUMENTS for all text after the command, $1/$2 for positional args. Lines starting with ! followed by a backticked shell command run and inline the output. @path references files. The body is the prompt Cursor runs when you invoke the command. Reference files with @path as needed.

Generated file 
 
 
 Where to save it 
 
 Project: 
  Personal: 
 
Project commands live in your repo and are shared with the team via
          git. Personal Claude Code commands stay on your machine.

 
 
 

Generation runs entirely in your browser. Command names, prompts, and
      tool settings never leave your device.

 
 

Slash commands are markdown files that turn a reusable prompt into a
        one-word shortcut. Claude Code reads .claude/commands/*.md
with optional frontmatter for descriptions, argument hints, and allowed
        tools. Cursor reads .cursor/commands/*.md as plain markdown
        with a title heading and prompt body.

 

Start with a prompt you already repeat: code review, release notes, test
        scaffolding. Add $ARGUMENTS where user input should land.
        Commit project commands so your team shares the same workflows.