That does seem to be what’s happening, and it appears to be the same problem faced in these other threads:
All of these reports come from the last two weeks, which makes me wonder if something has changed on GitHub’s side (or it could just be a coincidence).
I’ll gather these reports and try to get some attention internally. In the meantime, if it’s possible to close the pull request (in one of the other cases it was obviously an erroneous PR), that should solve the issue.