Track vulnerabilities like this one. Sign up free to get alerted when software you use is affected.

CVE-2020-28975: Libsvm in scikit-learn and other products can crash with a specially crafted model

CVE-2020-28975 · published 5 years ago

Summary

A bug in the svm_predict_values function in the Libsvm library, used in scikit-learn and other products, can cause a program to crash if it is given a specially created model. This could potentially happen if a user or another application manipulates the model in a way that the library is not designed to handle. To avoid this issue, users should ensure that they are not modifying the model in an unintended way.

What to do

Update scikit-learn developers scikit-learn to version 1.0.1.
Update scikit-learn scikit-learn to version 1.0.1 or later.

Affected software

Ecosystem	Vendor	Product	Affected versions
pip	scikit-learn developers	scikit-learn	>= 0.23.2, < 1.0.1 Fix: upgrade to 1.0.1
–	scikit-learn	scikit-learn	>= 0.23.2, < 1.0.1 `cpe:2.3:a:scikit-learn:scikit-learn::::::::`

Original advisory text

svm_predict_values in svm.cpp in Libsvm v324, as used in scikit-learn 0.23.2 and other products, allows attackers to cause a denial of service (segmentation fault) via a crafted model SVM (introduced via pickle, json, or any other model permanence standard) with a large value in the _n_support array. NOTE: the scikit-learn vendor's position is that the behavior can only occur if the library's API is violated by an application that changes a private attribute.

References

Fix within

Internet-facing 60 days

Internal At next upgrade

Not known to be exploited
Needs hands-on effort to exploit
Gives an attacker partial control

Severity

7.5 High

Exploitation

3% chance of attack within 30 days

Timeline

Published21 Nov 2020

Updated4 Nov 2026

First seen6 Mar 2026

Sources

GHSA-jxfp-4rvq-9h9m · GHSA

CVE-2020-28975 · NVD

Track software like this

Free during beta