Does a malicious proof need to exercise the same bug in all the engines necessarily? Why couldn't it exercise a different bug in each engine?
You could maybe address this through holding out the other implementations and using them only to verify?
You could maybe address this through holding out the other implementations and using them only to verify?