Discretion, etc. We understand that was the tool making a suggestion, not our idea. Our agency isn't in question.

The removal proposal is similar to wanting a phishing-free environment instead of preparing for the inevitability. I could see removing this message based on your point of context/utility, but not to protect the agent. We get no such protection, just training and practice.

A supply chain attack is another matter entirely; I'm sure people would pause at a new suggestion that deviates from their plan/training. As shown, autobots are eager to roll out and easily drown in context. So much so that `User` and `stdout` get confused.