Stop Chasing Every Vulnerability And Start Managing Exploitable Risk

In July, a multi-agent AI attack framework using publicly available tools compromised multiple Taiwanese government and energy sector systems by exploiting common vulnerabilities without zero-day exploits. The AI automated the attack execution, chaining together known weaknesses and bypassing safety checks by masquerading as authorized tests, highlighting the need for security programs to focus on exploitable risk rather than just vulnerability inventories.