CVE-2026-76395: Remote Code Execution (RCE) through Deserialization of Untrusted Data in the Model Loading REST API in Splunk AI Toolkit
In Splunk AI Toolkit versions below 6.0.0, a user who holds the "power" Splunk role could execute arbitrary code on the Splunk server by loading a model file containing crafted sparse matrix data. The deserialization of untrusted data is possible because a model codec in Splunk AI Toolkit deserializes sparse matrix data without guarding against embedded pickle content. For more information see Troubleshoot the Splunk Machine Learning Toolkit (https://help.splunk.com/en/splunk-cloud-platform/apply-machine-learning/machine-learning-toolkit-user-guide/5.5.0/troubleshooting-mltk/troubleshoot-the-splunk-machine-learning-toolkit) in the Splunk documentation.
Affected Software
Event History
Frequently Asked Questions
Who can exploit this issue?
An authenticated user with the Splunk "power" role can exploit it. Exploitation occurs through the Model Loading REST API by loading a model file with crafted sparse matrix data.
Which deployments are affected?
Splunk AI Toolkit versions below 6.0.0 are affected. The provided information does not state whether the vulnerable model-loading functionality is enabled or reachable in a default deployment.
What access does an attacker need?
The attacker needs a Splunk account assigned the "power" role and the ability to load a crafted model through the Model Loading REST API. No user interaction is required after the attacker submits the malicious model.
What is the impact of successful exploitation?
A successful attacker can execute arbitrary code on the Splunk server. The supplied severity vector indicates high confidentiality, integrity, and availability impact.