Attackers Abuse AntiFox AI Agents to Access AWS Secrets and Delete S3 Objects
Emerging Technology Risk · 14 sources · Updated 3d ago · First seen Nov 11, 2026

Attackers are abusing self-managed instances of AntiFox, an open-source AI-driven penetration-testing platform hosted on Alibaba Cloud, to compromise internet-facing web applications and associated AWS environments. Arctic Wolf Labs observed IIS compromise, ASPX web shell deployment, configuration and credential discovery, and cloud-service enumeration. Intruders used valid IAM user credentials and assumed EC2 roles to make authenticated AWS API calls, including successful Secrets Manager requests that exposed production secrets. A separate incident involved reading, copying, and deleting Amazon S3 objects.
The vulnerability used to compromise IIS and the source of AWS credentials in some attacks remain unknown. The activity illustrates how agentic tools can accelerate established intrusion techniques, but does not establish that every attack step was autonomous. Defenders should correlate IIS web shell activity with unusual AWS access, investigate successful API calls even when surrounded by denied requests, and revoke compromised sessions and credentials. Reviews should prioritize Secrets Manager access and S3 object operations to determine whether attackers exposed secrets or altered cloud data.


