@Wolf480pl@niu.moe
I considered that, and there are ways to address that to varrying degrees.
1) The most secure solution would work by each server in the federation spoofing a user on others servers int he federation. While this does still require the servers to cooperate (open registration mechanisms and such) it means that a server wouldnt be able to tell the difference between one of its own users and a spoofed user used to detect bans. As such even modifying their internal code wouldnt prevent detection.
2) work on a simpler good faith method. Basically if and when a server is found to violate the agreement the domain becomes permanently blocked. Otherwise trust the code is running as expected