Israeli firm Dream documents what appears to be the first autonomous AI attack on a government.
Cyera researchers discovered four OpenClaw vulnerabilities that form an attack chain allowing data theft, privilege escalation, and persistent backdoor access.