A Meta AI security researcher said an OpenClaw agent ran amok on her inbox 


Read on TechCrunch