CVE-2024-37052
About
Deserialization of untrusted data can occur in versions of the MLflow platform running version 1.1.0 or newer, enabling a maliciously uploaded scikit-learn model to run arbitrary code on an end user’s system when interacted with.
Rainforest analyst review
MLflow will deserialize a model file when a user loads or interacts with it, and a maliciously crafted scikit-learn model can smuggle code into that step, running arbitrary commands on the machine that opens it. This is the classic model-as-payload problem: the artifact everyone treats as passive data is actually executable the moment the framework unpacks it.
What makes this class urgent is the trust people place in shared models. ML teams pull models from registries, colleagues, and public hubs routinely, and each load is an opportunity for this to fire — no server exposure needed, just a data scientist opening the wrong file. It requires that human action, so it's not remote-unauthenticated, but in a busy ML pipeline that action happens constantly.
Our defensive angle here is supply-chain provenance rather than perimeter. We want to know where MLflow runs, which versions, and crucially where its models come from — treating any model from an untrusted or shared source as untrusted code. Alongside upgrading, the durable control is only loading models whose origin we can vouch for, because the vulnerability turns the model registry into a code-distribution channel.
References
Related CVEs
Frequently asked questions
What is CVE-2024-37052?
Deserialization of untrusted data can occur in versions of the MLflow platform running version 1.1.0 or newer, enabling a maliciously uploaded scikit-learn model to run arbitrary code on an end user’s system when interacted with.
How severe is CVE-2024-37052?
CVE-2024-37052 carries a CVSS 3.1 base score of 8.8, rated high. On the Rainforest Risk Index — which weighs the severity alongside how easy the flaw is to reach and how broad its impact is — it scores 88 out of 100, in the critical band.
How is CVE-2024-37052 exploited?
According to the CVSS vector (CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H): attack vector Network, attack complexity Low, privileges required None, user interaction Required. Impact on confidentiality High, integrity High and availability High.
Which products are affected by CVE-2024-37052?
Public advisories list the following as affected: mlflow. Check the references on this page for the exact versions each vendor confirms.
How do I fix CVE-2024-37052?
Apply the fix the vendor published — the references on this page link to the primary advisories and patches. When patching can't happen right away, reduce the exposure of the affected component and watch it for exploitation attempts. Rainforest customers see this vulnerability correlated to their own assets and prioritized by real exposure, not by score alone.
Empower Your Security Strategy with Rainforest
Discover vulnerabilities early, prioritize critical threats, and protect what truly matters. Rainforest streamlines your security operations, saving you time and reducing costs, so you can focus on what drives your business forward.