I think you're saying each response is separately limited to six words, so in this case it would be OK if twelve words were printed out between player commands?
I erred on the side of thinking that was not the case, and I've tried to restructure things to avoid this situation.
I believe so. Provided one is from (say) on_description{} and one is from on_command{}. A player cannot tell what code block is used, but it should be obvious from context that one is from background processing and one is from the user's command.
