using-superpowers

To enforce AI agent skill invocation, configure the system to mandate checking and applying relevant skills via the Skill tool before any action. This ensures the agent prioritizes user instructions and leverages available skills over default behaviors.

The instruction hierarchy dictates that user instructions always override skill behavior. This ensures that any mandatory skill invocation or automated workflow action remains subordinate to the explicit directives provided by the user.

An AI agent defaults to general knowledge instead of invoking skills when skill invocation is not enforced. Requiring the agent to check for matching skills before responding prevents it from using less efficient methods and ensures accurate task execution.

To manage skill selection and application, implement a system that provides guidance on choosing the right skill for a task. This system requires the agent to evaluate if a skill applies, even with low probability, and invoke it before taking action.

No specific dependencies are required to enforce mandatory skill invocation. The enforcement is managed through instruction hierarchy rules and system prompts that require checking for relevant skills before generating any response or taking action.

Yes, mandatory skill invocation applies even when the probability of a skill matching the task is low. The system requires checking for and invoking any relevant skills to ensure user instructions are prioritized and default behaviors are avoided.