As far as I know it has never been released publicly, but I did test out a shell annotation on our test server a few years ago.
I think it’s important to keep in mind that annotations serve two main purposes on iNat: for the photo browser, and for the graphs on the taxon page. When I tested out a shell annotation, it didn’t seem to fit either of those use cases. Pretty much any photo of a shelled mollusk includes the shell, so I didn’t see a use case for filtering by shell in the taxon photo browser, and any use for taxon page charts also seemed very limited.
This is all about actual utility, not about making sure every taxon has an annotation available for it. If it was the latter, we could have hundreds of annotations. If no annotation value is a good fit for an observation, that’s fine. Just don’t add an annotation.
That being said, I’m not a malacologist. If there’s a good case to be made for adding some sort of empty shell annotation that will meet the goals I mentioned above, I’m open to it.
The difference here is that, as I mentioned above, the shell of a shelled mollusk is pretty much always viewable in every photo (yes, I know there are some exceptions for internal shelled mollusks). Whereas with vertebrates the bones are often not, and being able to search for photos of just bones is helpful if someone comes across vertebrate bones and wants to identify them.
I’m happy to be corrected, but it seems like with both feather and fur, those are cases where a) for ID it’s useful to be able to search for photos of unattached feathers and tufts of fur and b) those are things that, when they come off of an animal, don’t cause the animal to die. Those seem fundamentally different to me than a mollusk shell.