dumb q alert: guessing this Inumber is always unique (after looking at the current code too)? Do we have anything like an Inode for filesystems in wnfs? https://en.wikipedia.org/wiki/Inode.

Yes, this number should be unique (unless there's a malicious peer with write permissions or a buggy implementation). You could think of files and directories at inodes. The term inumber also appears in the linked articles. It would then be the number that refers to a certain inode.
However, in our case the inumber is persistent across revisions. This is where the analogy kind-of breaks down. I'm not super sure whether inumbers change between revisions in traditional file systems, or whether everything basically happens via in-place mutation.

code-wise, do you see any aliasing needed for hardlinks now (vs symlinks)?

No. I don't think so.

I'm still thinking idly about the symlinking stuff in WNFS.
Only in the public file system hard-links correspond directly to CIDs.
In the private file system, that's not quite the case.

If our "own" version of hard-links doesn't become a thing, this will probably be a comment on the cp function of the public file system, as it doesn't actually need to copy anything, it's similar to a hard link.