Loving Sophia and its responsiveness. I also love that you can forward un-answered to questions to an LLM.
It would be great if that forward could be accompanied by a configurable prompt either that or, as suggested in another thread using a configured HA chat entity or alternative pipeline.
I expect 95%+ of requests I answer to be handled by Sophia, for the remainder that I goes of the big bad internet I'd love to provide guidance for how it should respond.