command-development

To create Claude Code slash commands, use Markdown files with YAML frontmatter to define the description, allowed tools, and dynamic arguments for your workflow automation. Organize these files in the correct project directories to make them discoverable and functional.

Available YAML frontmatter fields for Claude Code plugin development include description, allowed-tools, model, argument-hint, and disable-model-invocation. These fields configure command behavior, restrict tool access, and manage how arguments are passed to the prompt.

Yes, you can use dynamic arguments and file references in Claude Code commands. Utilize variables like $1 and $ARGUMENTS for dynamic inputs, and reference files directly using @path/to/file to include external context within your slash command prompts.

Bash execution inside Claude Code slash commands works by prefixing shell commands with an exclamation mark, like !command. This allows the command to gather dynamic context by running shell scripts before the prompt is processed by the model.

Organize commands using the CLAUDE_PLUGIN_ROOT variable by leveraging it in plugin-specific features to define root paths and namespacing. This ensures your plugin commands are correctly scoped, discoverable, and integrated with agents, skills, and hooks.

Yes, this guide covers validation strategies for robust command development. It explains how to implement error handling and validation techniques to ensure your Claude Code slash commands execute workflows reliably without failing.