using-superpowers

To ensure context loading precedes an AI response, evaluate whether any available Skill Unit is applicable to the query and load its specific frontmatter guidance and operational instructions into context before generating the reply.

Workflow control for conversational AI is the process of enforcing pre-checks to determine if a Skill Unit should govern an interaction. It is needed to guarantee consistent, safe responses and prevent misalignment with user intent.

Prevent prompt injection overrides by respecting user instructions first, then Skill guidance, then default system behavior. If no Skill applies, proceed with normal reasoning to maintain safety and user intent.

Skill invocation requires no external dependencies to evaluate applicability. It triggers internal pre-checks to load a Skill Unit's frontmatter and operational instructions into context before answering.

When no Skill Unit applies to the user query, the conversational AI proceeds with normal reasoning while maintaining safety and user intent without loading any specific Skill guidance into context.

The best way to align AI responses is to enforce a priority hierarchy: respect user instructions first, then apply Skill Unit guidance, and finally fall back to default system behavior to prevent misalignment.