Copied from discussion on Commons as it turned out that I am not alone with this request.

It would be great to have some special pages that would show:

• local project files with a duplicate on Commons
• local project file description pages shadowing a file description page on Commons
• local project files shadowing a Commons file where the local file is not a duplicate of the Commons file
• orphaned file talk pages on local project (where a file with the same title does not exist on Commons)

Now it takes some unpleasant hacks to get that information that is needed for projects’ maintenance.

For example, you can get a list of duplicates via API:
https://ru.wikipedia.org/w/api.php?action=query&generator=allimages&prop=duplicatefiles&gailimit=500&format=xml

But you get limits (no more than 5000 from bot account) and that seems inefficient/expensive wastage of resources and time to scan all files to get duplicates