News for hackers. Stuff that matters.
Re:Launch HN: April (YC S25) – Voice AI to manage your email... by svota on Monday August 25, 2025 @12:48PM (#45015858)
How are you handling the attack vector of in-context commands[1]?

[1]: https://guard.io/labs/scamlexity-we-put-agentic-ai-browsers-... (currently on the front page)

Re:Launch HN: April (YC S25) – Voice AI to manage your email... by dy5topian on Monday August 25, 2025 @02:06PM (#45016859)
that's a really interesting one
Re:Launch HN: April (YC S25) – Voice AI to manage your email... by danenania on Monday August 25, 2025 @09:00PM (#45021061)
I think you could probably solve it for this use case by just including a confirmation step for potentially destructive actions which is isolated from other context.