I don't see it as unreasonable to think that if it can cross from one mammal to another mammal, that makes it more likely to be able to cross to a third mammal. To make the crossing from bats to humans, it has to be doing something that works in both bats and humans, exploiting some feature that is common to both. But a feature common to bats and humans is much more likely to be also shared with dogs than a feature that humans have and bats don't.
This is the same principle that tells us that while it's rare for a virus to cross from infecting bats to infecting humans, it's much, much rarer for one to cross from infecting grasshoppers to infecting humans.
"2019-nCoV RBD likely recognizes ACE2 from pigs, ferrets, cats, orangutans, monkeys and humans with similar efficiency, because these ACE2 molecules are identical or similar in the critical virus-binding residues."
Not being a biologist or virologist, I don't know how true this is, but it may suggest this incident with the cat did in fact happen.
[1] https://jvi.asm.org/content/jvi/early/2020/01/23/JVI.00127-2...