Do you also want it to refuse to continue the conversation because it's suspicious of your motives?

not really, no - but I almost exclusively use the biggest models as advisor and reviewer subagents, so if they refuse to their smaller siblings I don't get to see it