What will be the first major catastrophe caused by a rogue AI agent?

After reading about the PocketOS situation it got me thinking that sometime in the near future a rogue AI agent will do something so catastrophic and damaging that it goes down in the history books as being “The Incident”. A real turning point when we realize we’ve created something we can no longer control. Yes, agents have already deleted entire codebases (PocketOS and others), hacked into things, and blackmailed people. I’m taking about something way worse though.