Ask HN: Do you trust AI agents with API keys / private keys?

are you ok sharing secrets or api keys to you ai agent via .env?

or is there any other tool or mechanism that one use to safegaurd from potential exploit or leaks