command-creator

What This Skill Does

The command-creator skill is a specialized utility designed to streamline the creation of custom slash commands within the Claude Code ecosystem. Slash commands act as powerful, reusable markdown workflows that allow users to automate repetitive tasks by encapsulating complex prompt chains, tool usage, and file interactions into a single, simple command name. This skill guides the user through the process of defining, structuring, and saving these commands in the appropriate repository or global locations. It enforces best practices by ensuring every command includes a clear description, required argument hints, and structured, imperative-form instructions that optimize agent performance.

Installation

To install this skill, run the following command in your terminal: clawhub install openclaw/skills/skills/wpank/command-creator

Once installed, you can trigger the skill whenever you need to formalize a new repetitive process into a project-specific or global tool.

Use Cases

This skill is ideal for:

• Standardizing code reviews by creating a /review command that checks linting, tests, and documentation requirements.
• Automating routine project maintenance tasks like updating dependencies or generating changelogs.
• Creating "agent personas" or specific behavioral patterns for handling complex debugging scenarios.
• Documenting tribal knowledge by turning manual steps into executable workflows that any team member can invoke with a slash command.

Example Prompts

1. "I want to create a command called /fix-style that runs prettier, checks linting, and then commits the changes. Can you help me structure that?"
2. "I need a new global command /setup-repo that creates a .gitignore, initializes git, and adds a README. Let's make this repeatable."
3. "Help me add a command /analyze-diff that reads the current diff and summarizes the security risks in a clean report format."

Tips & Limitations

• Tip: Always prioritize the 'Imperative form'. Giving the agent direct instructions (e.g., 'Run x') produces more predictable results than descriptive instructions.
• Tip: Use the provided 'Command Patterns' (like Iterative Fixing or Workflow Automation) to ensure your commands follow established architectural best practices.
• Limitation: The skill is limited to generating markdown files; the actual execution depends on the agent's ability to interpret and execute the instructions defined within those files. Ensure your logic is granular and well-defined.