Attackers can exploit the bugs through prompt injection, chaining them together to escape the sandbox and execute arbitrary code.
The post CrewAI Vulnerabilities Expose Devices to Hacking appeared first on SecurityWeek.