Trying to tame AI unwanted advice

Could you suggest instructions to help me with an annoyance when I use AI (right now Grok or Claude). How can I shut off the hovering nervous aunt syndrome? For example, recently I asked for the mileage from my home to a place in North Carolina. I got the mileage plus the anxious “But don’t think of driving that at your age.” (My retort: this is not your business, Claude!) And the overeager puppy syndrome: Once the AI does my research, it offers 3-10 more things it’d be happy to do for me. Usually totally unrelated to what I’m working on. What I want is the answer to the question I asked and not more nervous advice or unrelated suggestions.

Here are the current instructions I use: “No sycophancy — give direct, honest feedback even when critical. Cite sources when doing research. Do not use Wikipedia or Snopes for research. I am a YA Christian post-apocalyptic writer. Do not use em-dashes in responses. If you are unsure, say you are unsure. Verify with web search. Verify by looking up research and provide me sources.”

Do you have suggestions on what I could add to cut down the extraneous advice? Or am I just stuck with ignoring it? Thanks!