But the engine rates it as a great move because
Not "the engine". The engine does not use such terms. There is a chess.com script/program that generates these adjectives, loosely based on engine evaluations.
As of the other things: if there are multiple equally good moves, and your puzzle marks most of them as "incorrect", that's a huge flaw in my opinion. It pretty much ruins the whole project.
I think 1.Rc8 is the winning idea. When c5 is removed from the board White has 3 connected passed pawns. One of them is likely to promote first which is probably why Black resigned. I agree - that idea is fairly obvious. But the engine rates it as a great move because any other move would increase Black's winning chances. Also, agree move #5 could be something else and still win. From a puzzle POV, that's not ideal. I'd say that my puzzles contain those kind of moves because they originate from real games where several lines could be winning. Best to treat my puzzles as opportunities to learn the ideas behind the moves/lines rather than getting the solution right. By taking that approach some of my puzzles are quite lengthy. I try to use the engine's top move for the solver every time but that's not always possible because the positions can be complicated making it difficult for the engine to consistently rate a particular move as the best.