Some questions about S3

I saw that rocketmq connect was designed as a tool for streaming data across heterogeneous systems. But I have a few questions, as follows:
1、Now, the connectors support only partial heterogeneous systems of data processing? Because I have looked at some of the code related to sink task in the master branch, such as (sftp, cassandra)sink task, their source/sink task is processed in a special data way. It may not support things like redis source -> sftp sink this way.
2、SSo, for S3, need to support S3 source -> S3 sink this way? Or do you need to support all heterogeneous systems?
3、If need to support other heterogeneous systems, like S3 sink for batch message processing such as Mysql source, it needs to generate table files?