I am a little confused on why the author makes a distinction between Cellebrite using zero-days to hack a phone to read data and Signal's hack. While the US government might have the framework to not be considered violating CFAA, what about when other governments use Cellebrite? From this point of view, Cellebrite isn't a valve that's stopping back door decryption in systems, it is the back door. Signal including these files is in a sense covering a back door.

Also, is leaving a file that breaks the admissibility of previously gathered evidence considered active hacking? Am I misunderstanding something about the function of the files in Signal? I thought the only way Cellebrite's software is interacted with is if it tries to access Signal on the device. Signal isn't actively searching to hack back. It's triggered by Cellebrite's software, not Signal's.

The straightforward workaround would be to delete Signal before using the Cellebrite software which I think is the real point. Signal isn't trying to protect the end user actively and can't do anything if it's not installed on a phone.