Regarding the first two, it certainly is possible to do file-based operations with Meltano, but most users pull from a source system (SaaS or database) and load to a destination system (most often a database or data warehouse).
Do you have anything upstream from the jsonl source that can be queried directly? And for the file-copy operations (s3-to-s3 or s3-to-gcp), do you only need the files copied or do you also want rich metadata that you'd get by running as an extract-load operation?