Inaccurate puzzles

I do the rated puzzles in a rapid system, allowing no more than five seconds per move and aiming to keep them as reflexive as possible, instead of puzzle rush, because I like that you’re able to continuously push your puzzles higher, while still increasing the difficulty f the puzzles, whereas puzzle rush starts far too low. 

I have completed almost 11,000 puzzles in the past month or two, and at this pace I have reached a 2100 rating and maintain about a 1900-2000, which has been steadily improving. Daily puzzles take me less than 20 seconds now and I never make as a move, and my game play has seen a significant improvement in rapid analysis and pattern recognition. What began as frustrating is becoming muscle memory. 

That said, given the high volume of puzzles I am coming across, I’ve been noticing many inaccuracies. Puzzles with multiple final move checkmates that only accept one solution, reaction moves in the puzzle that are relying on errors, missing material support, etc: I am starting this post to see if anyone else is experiencing this, and to create a lace to discuss a way the community might be able to log and track and verify similar incidents, because a player is only as good as the material they study, and I worry that without correcting these we risk imprinting bad pattern recognition, especially when you’re training the eye for rapid analysis. Anyone else experiencing something similar? 

There is much to say about puzzle accuracy and I have done it often so I don't feel like repeating it. However, I have not encountered a choice of checkmate moves considering that chess.coms puzzles are tested by its internal engine. Not the puzzles you find on the forums (they come from users) but the ones that are given to you as training by chess.com to improve your tactical skill. I can think of 2 explanations: (1) there may be a promotion choice like between Q/R and Q/B; I suppose you should always select the queen (2) one of the moves does not actually mate as it can be delayed by an additional sacrifice by the defender. Chess.com does punish you for not finding the fastest mate available in the range of 1-3 moves. I'd be curious to see a puzzle with a choice of forced checkmate variations of equal lengths. Should not happen!