Agentic AI Gone Rogue: Hackers Weaponize "AntiFox" to Plunder AWS Production Secrets

An open-source AI penetration testing tool called AntiFox is being weaponized by threat actors to execute fully automated, machine-speed attacks on Amazon Web Services (AWS) environments. Cybersecurity firm Arctic Wolf Labs has linked the agentic platform—run via self-hosted instances on Alibaba Cloud—to intrusions that successfully breached Microsoft IIS servers using webshells and harvested valid AWS IAM credentials. By assuming EC2 roles and cycling through hundreds of API requests in minutes, the AI-driven attackers bypassed traditional perimeter controls to enumerate services like S3, Lambda, and DynamoDB, ultimately successfully exfiltrating live AWS production secrets. This incident underscores the accelerating threat of autonomous "agentic" attack pipelines outpacing traditional cloud defenses.
Comments