using-superpowers

To ensure an AI agent invokes skills before responding, enforce a tool-first interaction model. Mandate the use of the Skill tool for any task where a skill might apply, preventing unassisted actions and promoting a structured, tool-augmented workflow.

An LLM workflow generates answers without checking available skills because it lacks a mandatory skill invocation step. Applying a structured enforcement ensures the AI prioritizes skill discovery and execution over immediate response generation for all conversational turns.

A tool-first interaction model for AI agents is a structured workflow that mandates leveraging available tools and skills before generating any response, including clarifying questions. It applies to all turns where a skill's applicability is even slightly probable.

Enforce mandatory skill invocation across all conversational turns by mandating Skill tool usage whenever a skill's applicability is slightly probable. This guides the AI to first invoke the relevant skill to gather information before formulating a direct answer.

You need to enforce skill discovery in an AI agent workflow whenever a user asks a question that could potentially be answered or assisted by a known skill. This prompts the AI to first invoke the relevant skill to gather information.

Yes, mandatory skill invocation applies to clarifying questions. The enforcement ensures AI agents consistently leverage available tools and skills before generating any response, preventing unassisted actions and applying structured tool-augmented workflows.