While I am no expert on Xenophoridae (carrier shells), it has always been my
impression that the "glue" they use to attach other shells to their own shells
is . . . shell.  As I understand it, once they get a new "attachment"
correctly positioned on their shell (and heaven knows how they do that!),
their mantle (the organ that produces their own shell) secretes a thin layer
of new shell material into the gap between their shell and the intended
attachment.  Once they have it "tacked" in position this way, they proceed to
secrete additional layers of shell material, filling in more and more of the
space under and around the attachment, until finally the part of the
attachment in contact with their shell is surrounded by, or embedded in the
material of their own shell.  In the all-to-common event that you bump one of
the attachments on your carrier shell specimen and knock it loose, it is often
easy to reattach it with little glue to it in its exact original position,
because the attachment, when removed, often leaves an exact imprint in the
carrier shell, into which the attachment fits very precisely, thereby allowing
precise repositioning.
 
Paul M.