CVE-2026-76393: Race Condition during Model Upload through the REST API in Splunk AI Toolkit
In Splunk AI Toolkit versions below 6.0.0, a user who can upload models could overwrite a model being uploaded by another user by sending a concurrent upload request for the same model name, causing the resulting model lookup entry to reference attacker-controlled content. The race condition is possible because Splunk AI Toolkit does not verify that the uploaded content belongs to the request that creates the model lookup entry. 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 attacker needs permission to upload models through the REST API. They must also be able to submit a concurrent upload using the same model name as another user's upload.
Does exploitation require user interaction or special timing?
Yes. Exploitation depends on winning a race between concurrent model-upload requests, and the supplied vector indicates user interaction is required. The attacker attempts to cause the model lookup entry created by the other upload to reference attacker-controlled content.
Which versions are affected?
Splunk AI Toolkit versions below 6.0.0 are affected. The provided information does not state whether any particular default configuration enables or restricts model-upload access.
How can I tell whether this may have occurred?
Review model uploads and model lookup entries for cases where a lookup entry references content not belonging to the request or user expected to create it. Focus on concurrent uploads that used the same model name.