the Content Segmentation step isn't finding a "diff" of the two databases. It's importing the whole source database into the target, causing loads, redundancy, and unnecessary bloat in the migration files. It should use a more sophisticated diffing process
it's not meant for bidirectional usage right now. it assumes the lower environments' installations are nested in that of the production environment. theoretically, syncing staging to development would work, but production to staging would not. need to optimize this logic, as syncing production with lower environments could be useful, though it is outside of the project's original scope