Researchers warn that a newly identified open-source AI security testing platform called CyberStrikeAI was used by the same threat actor behind a recent campaign that breached hundreds of Fortinet FortiGate firewalls.