I'll take a look at it.
It is definitely possible to embed metadata in rtf files. I have not yet looked at what programs include what.
Going from the current Microsoft spec it allows at the very least for an "Information Group" (Page 38 in the document).
This includes a title, subject, author, company, several timestamps and others.
Stripping all \info tags comes to mind, but title and keywords might be desirable to keep.
I just skimmed the document and have not tested Word/LibreOffice/OpenOffice what they include. So there might also be other metadata embedded apart from the Information Group structure
The keywords are definitively metadata, and I think that the title might be considered to be one too: users would be less surprised if find that mat2 removed the title, than the other way around.
Another thing is pictures. PDF handling (according to doc/implementation_notes.md) is done by basically converting it to a PDF picture (default) or ignoring metadata of embedded objects (lightweight cleaning).
Since RTF doesn't allow for too many types of embedded content I think cleaning these is feasible.