So is it always accurate to label a question as "tricky"?

If every question is tricky, then the label of "tricky" ceases to be meaningful.

But it occurred to me that if the question had been identical to one given in class but I still labeled it as "tricky" the "tricky" label would have been accurate because the trick would have been students thinking that the obvious answer wasn't correct when indeed it was.

I believe the word you're looking for is "cruel".