Either you are pretending you don't understand what attack vector I am talking about, or you don't understand how the software you are actively defending works under the hood. The "official" Wasabi Coinjoin coordinator doesn't perform their own blockchain analysis: it doesn't know in advance whether certain UTXOs are good or bad: it might be those attempting to ruin the CoinJoin process or those attempting to ruin the company's reputation. The latter type of UTXO is evidently more dangerous since it directly affects the company's well-being and income. Criminals, dissidents, opposition parties, journalists who dared to uncover inconvenient truths, and others are now considered sources of "naughty" coins: they should be detected early and denied mixing service, which is why the company begins sharing part of its income with "professionals" supposedly capable to tell bad coins from not-so-bad ones effectively. The company and its coordinator blindly rely on data coming from external closed-source API, and they have to trust everything they are told because of the design of CoinJoin software. When parts of your open-source software rely on closed-source sources of information, it can no longer be considered fully open-source and trustless: both a coordinator and its customers now depend on the blockchain surveillance firm not behaving maliciously. Coinfirm, a blockchain analysis company, acting in their own interests and approving only those coins they want to deanonymize is a potential attack vector that cannot be prevented or detected in advance.