CVE Catalog

CVE-2026-41486

High
Published: Translated: NVD NIST

Summary

Ray is an AI compute engine that, from version 2.54.0 to before 2.55.0, registered custom Arrow extension types in PyArrow. As a result, when reading a Parquet file containing these types, arbitrary code execution is possible due to improper handling of metadata.

Risk Assessment

Organizations may be exposed to arbitrary code execution, potentially leading to data loss or system compromise. Specifically, an attacker could exploit this vulnerability for remote code execution.

Recommendation

It is recommended to upgrade Ray to version 2.55.0 or later to mitigate this vulnerability. Additionally, monitor and audit the use of Parquet files in the system.

Related vulnerabilities

Original NVD description (English source)

Ray is an AI compute engine. From version 2.54.0 to before version 2.55.0, Ray Data registers custom Arrow extension types (ray.data.arrow_tensor, ray.data.arrow_tensor_v2, ray.data.arrow_variable_shaped_tensor) globally in PyArrow. When PyArrow reads a Parquet file containing one of these extension types, it calls __arrow_ext_deserialize__ on the field's metadata bytes. Ray's implementation passes these bytes directly to cloudpickle.loads(), achieving arbitrary code execution during schema parsing, before any row data is read. This issue has been patched in version 2.55.0.

Vulnerability data from NVD (NIST) · CISA KEV · EPSS