Findings

CVE-2026-33744

BentoML is a Python library for building online serving systems optimized for AI apps and model inference.

Published Mar 27, 2026 · updated Oct 7, 2026 · bentoml bentoml

BentoML is a Python library for building online serving systems optimized for AI apps and model inference. Prior to 1.4.37, the field in accepts arbitrary strings that are interpolated directly into Dockerfile commands without sanitization. Since is semantically a list of OS package names (data), users do not expect values to be interpreted as shell commands. A malicious achieves arbitrary command execution during / . Version 1.4.37 fixes the issue.

Code

The project this record applies to. Exploit listings stay in References.

No public source repository is identified for this product.

Timeline

Detection, updates, and resolution for this record.

  1. Mar 27, 2026 · NIST NVD

    Detected

    BentoML is a Python library for building online serving systems optimized for AI apps and model inference. Prior to 1.4.37, the field in accepts arbitrary strings that are interpolated directly into Dockerfile commands without sanitization. Since is semantically a list of OS package names (data), users do not expect values to be interpreted as shell commands. A malicious achieves arbitrary command

  2. Oct 7, 2026 · NIST NVD

    Updated

    BentoML is a Python library for building online serving systems optimized for AI apps and model inference. Prior to 1.4.37, the field in accepts arbitrary strings that are interpolated directly into Dockerfile commands without sanitization. Since is semantically a list of OS package names (data), users do not expect values to be interpreted as shell commands. A malicious achieves arbitrary command

Affected

  • cpe bentoml/bentomlversion unspecified

Sources

  • CVE.orgVulnerability recordCVE-2026-33744

    Official CVE identifier

    cve.org
  • NIST NVDVulnerability recordCVE-2026-33744

    BentoML is a Python library for building online serving systems optimized for AI apps and model inference. Prior to 1.4.37, the field in accepts arbitrary strings that are interpolated directly into Dockerfile commands without sanit

    nvd.nist.gov

References