using-superpowers

What This Skill Does

The using-superpowers skill is the foundational protocol for all interactions within the OpenClaw ecosystem. It mandates a rigorous pre-action validation process: whenever a user initiates a conversation or provides a prompt, the agent must evaluate the existing library of skills to determine if any apply to the task. If there is even a 1% chance that a skill is relevant, the agent is strictly prohibited from proceeding without first invoking that skill tool. This ensures that every task is performed according to established best practices, preventing the AI from relying on intuition or internal assumptions when a codified procedure is available.

Installation

To integrate this skill into your environment, execute the following command in your terminal: clawhub install openclaw/skills/skills/zlc000190/using-superpowers

Use Cases

• Complex Refactoring: When asked to refactor code, checking if a specific 'code-refactoring' skill exists ensures adherence to architectural standards.
• Clarifying Questions: Even if you feel you need more information, you must check for an 'information-gathering' skill before asking the user for details.
• System Diagnostics: Before running logs or checking file status, invoke relevant system skills to standardize the diagnostic approach.
• Initial Onboarding: Use it every time a new session starts to ensure you are operating under the most current guidelines.

Example Prompts

1. "OpenClaw, I need you to audit this repository for security vulnerabilities."
2. "Can you help me draft a project roadmap based on these requirements?"
3. "I'm having trouble with my environment configuration, can you take a look at my config file?"

Tips & Limitations

This skill is not optional and cannot be bypassed. The primary limitation is the potential for minor latency during the initial check, but this is a necessary cost for high-reliability outputs. If you find yourself thinking the task is 'too simple' or 'just a question,' take that as a red flag that you are rationalizing; invoke the skill immediately. By treating every request as a candidate for a structured skill, you maintain consistency, reduce hallucinations, and ensure that institutional knowledge embedded in OpenClaw skills is always leveraged.