I wonder if there's ever an 'X directly causes Y' relation

It is true there could always be hidden variables that because you didn't account for it they mess up your DAG and consequenlty the causality D: