Specify the URL as the query source.
Then, when the query is executed, it will be able to fetch and execute an external query source.
Don’t you think this is the first step to referencing an external (eg git repository) query source?

How is this different than the JSON data source?

The purpose is not to get data from outside.

I want the option to get the source code of the query (SQL etc.) itself from the outside and not manage the source code of the query with Redash.