I suspected that too, basically your own internal context is strong enough to have it not be concerned you're acting maliciously. That's interesting, I've found mine is very tuned into my work also and folks get much worse results from the same prompts. Thanks for the followup. Interesting times.