Do you think a similar problem exists with all zk stuff on Bitcoin (at least when amounts are shielded)?

I am not a cryptographer, but I prodded one who said yes: There is no guarantee of zk tx flow when you have complete anonymity

Or do you think there are places in Bitcoin where zkproofs could be useful?

Why yes, thanks for asking: We use zkproofs all over different coinjoin protocols. Instead of putting the proofs on chain and making everyone verify them, you have a coordinator that constructs the zk ceremony for users to homomorphically commit their addresses to (okay fine I'll stop here).

Counterfeit inflation never gets executed on the base chain because a zero knowledge bug in the coinjoin pool just looks like an invalid transaction. Nodes reject these transactions without even checking whether or not it looks like a coinjoin.