Apply a named Thinking Toolkit model or audit an argument with /logic. Use only when the user explicitly invokes the toolkit, names an included model or alias, or invokes /logic.