Thanks, this was enough to track it down.
I found two issues. Your OPML uses xmlURL, and NewsBlur was only looking for xmlUrl, so those entries could be skipped. Also, a single feed failure could stop the rest of the OPML import.
I have a fix ready for both. NewsBlur now accepts xmlURL, and if one feed fails during import, it logs that feed and keeps importing the rest of the file.
The login required bsky feeds still will not import unless they are publicly accessible, but they should no longer break the whole batch.