Arbitrary Code Execution Vulnerability in MLflow Platform
CVE-2024-37058
8.8HIGH
What is CVE-2024-37058?
The MLflow platform version 2.5.0 and above is susceptible to a vulnerability that allows deserialization of untrusted data. This issue arises when a Langchain AgentExecutor model is maliciously uploaded, enabling attackers to execute arbitrary code on the systems of users interacting with such models. As a result, this vulnerability could potentially compromise system integrity, allowing unauthorized access and manipulation of system resources.
Affected Version(s)
MLflow 2.5.0