Ideally you would have some change tracking system (cdc f.e.) on the source tables (Streams in the case of Snowflake, Introduction to Streams | Snowflake Documentation).
But that is not the case.
So I think you approach is ok.  You cannot track what is not registered.
All you can do is do some kind of comparison between source/target.

About the cloning feature: if something is in public preview, it will become GA and is already stable.
So no issue in using it.