admittedly I am not an expert on bitcoin convenants. I know that there are alot of good use-cases for them. that said, how is allowing a utxo to be encumbered, especially in the case of recursive covenants, not a huge attack surface ripe for abuse? it sounds like a great way to make the bitcoin network an ofac-compliant cbdc.
Discussion
Replace covenants with multisig and realize that these 'attack surfaces' already exist. With covenants, the encumberance is coded into the recieve address, which you create and issue to the sender. The sender cannot suprise you with being in a covenant. Same with multisig, they cannot force you to withdraw into a multisig, but ofac compliant cbdc coin is possible (and easier) by only allowing withdraws into a multisig, whose second signer would only sign txn after checking with a gvt sponsored allow/block list... No covenant required
not the same thing though. in a multisig, I'd need someone else to sign off on my spending. with a covenant, I don't need permission to send but I'm still restricted on where it goes. in terms of censorship, that scales much better for the censor than sending funds to a multisig. plus I don't think "we can already do that" is a good argument for doing another bad thing. "I already stole. so what's a another theft?"
The difference is, the covenant restrictions you decide for yourself when you generate the recieve address. Nobody but you can restrict the future spends....
that would make sense. it still has the issue of being more easily scalable for this abuse. with a recursive covenant, the encumberance stays with the utxo as it's spent. as I understand it, you end up in a tainted Bitcoin scenario. at least the "tainted" Bitcoin we have now can have their history severed with a coinjoin