Bulletin ID: 2026-181-AWS
Scope: AWS
Content Type: Important (requires attention)
Publication Date: 11/12/2026 10:45 PM PST

Description:

AWS SDK for pandas (awswrangler) is an open-source Python library that integrates pandas with AWS services, including Amazon EMR.

We identified CVE-2026-127119, an issue in the EMR module of AWS SDK for pandas (awswrangler) from version 1.1.0 before 3.18.0. When a cluster is created without an explicit log location, the library uses a predictable default S3 bucket name for the cluster logs without verifying that the caller's account owns that bucket. A remote actor who creates that bucket first could read the cluster's logs, including job step output.

Impacted versions: >= 1.1.0 AND < 3.18.0

Resolution:

This issue has been addressed in awswrangler version 3.18.0. We recommend upgrading to the latest version and ensuring any forked or derivative code is patched to incorporate the new fixes. Users who rely on the default log location should also make sure the caller has s3:ListBucket on the default log bucket, which version 3.18.0 requires to verify bucket ownership.

Workarounds:

If you cannot upgrade immediately, pass an explicit log location that points to a bucket you own when creating clusters with AWS SDK for pandas, or create the default EMR log bucket (aws-logs-{account}-{region}) in your own account in each region you use. If that bucket name is already taken by another account, treat it as untrusted and use an explicit log location.

References:

• CVE-2026-127119 

Acknowledgement:

We would like to thank Upwind Security for collaborating on this issue through the coordinated vulnerability disclosure process.

Please email aws-security@amazon.com with any security questions or concerns.