Metadata cleanup

So, I once had a pile of Custom Metadata in a project, and then I discovered it wasn’t helping me. So I deleted the items.

However, on looking through the .Scrivx file (for other things), I noticed the data had stuck around on at least one entry (actually, 30+ entries). This surprised me.

Now, I can clean up the .Scrivx file, but I’m wondering why deleting the metadata doesn’t clean up the entries? The custom metadata list was gone, but 30 or 40 documents have this old metadata hanging around in the XML. Could this cause me issues down the line if I chance to re-use a fairly common field name? (“POV”, as an example) Hmm. The answer is yes. If I add a field with the same name as those old pieces of metadata, Scriv reuses those old entries.

Is this intentional? How do I clean up that stuff without editing the .Scrivx file?