I don't really know but maybe if you insert URL of <Clone Document> so the user clone some well-crafted Document, this document may access some other documents and a screenshot may leak them.

It only a possibility but usually once you have the XSS puzzle piece, getting the data may be as trivial as some JS code