CVE-2026-1777 affects the Amazon SageMaker Python SDK versions before 3.2.0 and 2.256.0. The vulnerability is due to the ModelBuilder HMAC signing key being transmitted in cleartext in the DescribeTrainingJob API response. This exposure allows an attacker with appropriate permissions to invoke the API and modify the Training Jobs S3 output location by uploading arbitrary artifacts. These artifacts could then be executed during later training job invocations, potentially compromising the training process and associated data. The CVSS 3.1 base score is 7.2, reflecting network attack vector, low attack complexity, high privileges required, and high impact on confidentiality, integrity, and availability. The vulnerability is categorized under CWE-319 (Cleartext Transmission of Sensitive Information). No patch or vendor advisory is currently available to confirm remediation status.

The vulnerability allows an attacker with permissions to call the DescribeTrainingJob API and modify S3 output objects to inject arbitrary artifacts that execute during training jobs. This can lead to unauthorized code execution within the training environment, compromising confidentiality, integrity, and availability of the training process and potentially sensitive data. The exposure of the HMAC signing key in cleartext increases the risk of credential compromise and unauthorized actions within the affected SDK environment.

Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. Until an official fix is available, restrict permissions to the DescribeTrainingJob API and S3 output locations to trusted users only. Monitor and audit access to these resources to detect unauthorized modifications. Avoid using vulnerable versions of the SageMaker Python SDK in sensitive environments if possible.