The IMDS impersonation issue involves the risk that AWS tools running outside EC2 instances on networks controlled or influenced by a third party may receive credentials from a maliciously impersonated IMDS endpoint. Normally, IMDS operates on a loopback interface within EC2 instances, ensuring metadata and credentials are securely delivered within the AWS data perimeter. This vulnerability arises when AWS CLI/SDK or SSM Agent are used on non-EC2 compute nodes in environments where an attacker can position themselves to respond to IMDS requests, potentially causing the client to interact with unexpected AWS accounts. The issue affects both IMDSv1 and IMDSv2. AWS advises customers to adhere strictly to installation and configuration best practices for these tools outside AWS environments and to monitor network traffic for IMDS-related endpoints and metadata requests to detect anomalous activity.

If exploited, this issue could lead to AWS tools outside the AWS data perimeter receiving credentials from unauthorized IMDS endpoints controlled by third parties. This may cause customers to unknowingly interact with unexpected AWS accounts, potentially leading to unauthorized access or data exposure. However, exploitation requires the attacker to have a privileged network position on the same network as the affected compute node. The severity is assessed as low by AWS, reflecting limited risk under typical deployment scenarios.

AWS recommends that customers using AWS CLI, SDK, or SSM Agent outside the AWS data perimeter strictly follow the official installation and configuration guides to mitigate this issue. Additionally, organizations should monitor their on-premises networks for unexpected IMDS traffic, including connections to the IPv4 link-local endpoint 169.254.169.254, the IPv6 endpoint fd00:ec2::254, HTTP requests to AWS metadata paths (/latest/meta-data, /latest/api/token), and the presence of AWS metadata headers such as X-aws-ec2-metadata-token. Detection rules, such as those provided in SIGMA format, can be deployed in SIEM systems to identify potential impersonation attempts. No official patch is indicated; mitigation relies on proper configuration and proactive monitoring.