Currently, various checks kill message processing before the bot even gets to forming a reply to a ping. I suggest moving the moderation check to be within each possible response type to a message, and remove some earlier checks.
I realize, by classifying the type of response required before doing a moderation check, we can also avoid wasting credits on messages that do not prompt an LLM response.
Currently, various checks kill message processing before the bot even gets to forming a reply to a ping. I suggest moving the moderation check to be within each possible response type to a message, and remove some earlier checks.