using-superpowers

To make an LLM agent invoke skills before responding, you enforce mandatory skill utilization using a dedicated tool, ensuring the agent checks relevant workflows before generating any answer.

AI agents skip tool use and provide incomplete responses when mandatory skill invocation is not enforced, allowing them to guess answers instead of checking predefined workflows and relevant tools first.

The best way to prioritize skills in an agent workflow is to apply a strict selection logic that prioritizes process skills over implementation skills for optimal task execution.

Yes, you need to configure skill invocation for every conversational turn where a skill might be applicable, ensuring comprehensive tool utilization and strict adherence to predefined workflows.

Yes, you can use mandatory skill invocation to prevent clarifying questions by forcing the agent to check and use relevant knowledge base or tool skills first rather than guessing.

If an agent selects the wrong skill type during tool invocation, it violates strict skill priority and type guidelines, resulting in suboptimal task execution and degraded workflow adherence.