using-superpowers

To enforce a skill-first workflow, you must configure the AI to evaluate and invoke the appropriate Skill tool before generating any response. This ensures every interaction starts by checking Skill applicability to guide behavior through explicit boundaries.

Misaligned responses occur when the AI lacks a pre-response Skill invocation workflow. Directing the AI to locate and apply the correct Skill tool before answering prevents misalignment by enforcing explicit Skill boundaries upfront.

The best way to apply tool invocation policies across every conversation is enforcing a Skill-first workflow. This requires a root SKILL.md with a valid name and description to load Skill content and evaluate applicability before the AI responds.

Yes, you can use a root SKILL.md file to direct AI clarifying questions and policy adherence. It enforces a pre-response Skill invocation workflow, ensuring the AI evaluates applicability and applies the correct tool before answering.

Yes, you need a root SKILL.md with a valid name and description to load Skill content. This is required for the AI to evaluate applicability upfront and determine the best course of action before responding in any conversation.

The workflow requires a root SKILL.md with a valid name and description to load Skill content. Without this prerequisite, the AI cannot evaluate applicability or enforce the pre-response Skill invocation, preventing the enforcement of explicit Skill boundaries.