using-superpowers

To enforce AI tool invocation before responses, establish strict workflow rules that prevent rationalization bypassing skill checks. This mandates checking applicable skills prior to output generation, ensuring the agent invokes relevant process or implementation skills first.

An LLM bypasses available skills for simple questions due to missing strict workflow enforcement. Without mandatory invocation rules, the AI rationalizes skipping tool checks, assuming unassisted answers are sufficient instead of applying relevant process skills.

Mandatory skill invocation in AI agent workflows is a strict enforcement mechanism requiring the agent to check and use applicable skills before answering. It prioritizes process and implementation skills based on task type, preventing unassisted actions.

To stop AI from rationalizing out of using skills during context gathering, establish strict rules against bypassing skill checks. Mandate tool invocation even for low-probability matches, preventing the agent from skipping skills when more context is needed.

An AI agent should prioritize process skills over implementation skills based on the specific task type. Establishing skill priority guidelines ensures the agent selects the correct tool-driven workflow for effective task execution.