CVE-2026-31237: Critical severity pypi/ludwig vulnerability

Q: Who is exposed to this issue?

Systems using the Ludwig framework through version 0.10.4 are exposed when they run predict() on a dataset path that an attacker can control or influence. The vulnerable path is reached when the supplied dataset is identified as a .pkl file.

Q: What does an attacker need to exploit it?

An attacker needs to cause the Ludwig predict() method to load a maliciously crafted pickle dataset file. No authentication, user interaction, or special attack conditions are indicated by the supplied CVSS vector.

Q: What is the impact of successful exploitation?

Loading the malicious pickle can deserialize arbitrary Python objects and lead to arbitrary code execution on the system performing the Ludwig prediction. The reported CVSS vector also indicates high confidentiality, integrity, and availability impact.

Q: How can I determine whether a workflow is affected?

Review Ludwig prediction workflows for calls to predict() that accept dataset file paths, especially paths sourced from users, external systems, shared storage, or downloads. A workflow is affected if it can pass a .pkl dataset to predict() on Ludwig version 0.10.4 or earlier.

Published May 12, 2026

Updated

The Ludwig framework thru 0.10.4 is vulnerable to insecure deserialization (CWE-502) through its predict() method. When a user provides a dataset file path to the predict() method, the framework automatically determines the file format. If the file is a pickle (.pkl) file, it is loaded using pandas.readpickle() without any validation or security restrictions. This allows the deserialization of arbitrary Python objects via the unsafe pickle module. A remote attacker can exploit this by providing a maliciously crafted pickle file, leading to arbitrary code execution on the system running the Ludwig prediction.

Affected Software

1 affected component

pypi/ludwig<=0.10.4

Event History

May 12, 2026

CVE Published

via MITRE·12:00 AM

Data Sourced

via MITRE·12:00 AM

Description

Data Sourced

via NVD·06:16 PM

DescriptionSeverityWeakness

Frequently Asked Questions

Who is exposed to this issue?

Systems using the Ludwig framework through version 0.10.4 are exposed when they run predict() on a dataset path that an attacker can control or influence. The vulnerable path is reached when the supplied dataset is identified as a .pkl file.

What does an attacker need to exploit it?

An attacker needs to cause the Ludwig predict() method to load a maliciously crafted pickle dataset file. No authentication, user interaction, or special attack conditions are indicated by the supplied CVSS vector.

What is the impact of successful exploitation?

Loading the malicious pickle can deserialize arbitrary Python objects and lead to arbitrary code execution on the system performing the Ludwig prediction. The reported CVSS vector also indicates high confidentiality, integrity, and availability impact.

How can I determine whether a workflow is affected?

Review Ludwig prediction workflows for calls to predict() that accept dataset file paths, especially paths sourced from users, external systems, shared storage, or downloads. A workflow is affected if it can pass a .pkl dataset to predict() on Ludwig version 0.10.4 or earlier.

CVE-2026-31237: Critical severity pypi/ludwig vulnerability

Affected Software

Event History

Frequently Asked Questions

Who is exposed to this issue?

What does an attacker need to exploit it?

What is the impact of successful exploitation?

How can I determine whether a workflow is affected?

Company

Resources

Contact

Frequently Asked Questions

Who is exposed to this issue?

What does an attacker need to exploit it?

What is the impact of successful exploitation?

How can I determine whether a workflow is affected?