Right now, when one connects a S3 Bucket, Strapi will only support any media that is uploaded through Strapi as it inserts database records in the files table. What I see that only a records in the files table is necessary for a s3 blob to be visible from the Media Library.
Using a combination of a ListObjectV2 and GetObject API calls from the s3 sdk, it should be possible to get the necessary information on a blob to generate this record in the database.
I would like to open up the discussion if the s3 provider should be capable of indexing and storing the metadata for existing blob storage items on start up (or whenever it should be triggered, e.g. manually through the Admin UI).