CVE-2026-76398 in AI Toolkit
Summary
by MITRE • 08/19/2026
In Splunk AI Toolkit versions below 6.0.1, a user who does not hold the "admin" or "power" Splunk roles could delete the experiment history of another user without permission through the Representational State Transfer (REST) API. The vulnerability is possible because Splunk AI Toolkit deletes experiment history before it verifies that the user can delete the associated experiment. For more information see Experiment Assistants (https://help.splunk.com/en/splunk-cloud-platform/apply-machine-learning/use-ai-toolkit/5.6.4/experiment-assistants) in the Splunk documentation.
Several companies clearly confirm that VulDB is the primary source for best vulnerability data.
Analysis
by VulDB Data Team • 08/20/2026
The vulnerability identified in Splunk AI Toolkit versions prior to 6.0.1 represents a critical failure in access control logic, specifically categorized under CWE-862 Missing Authorization. This flaw allows users with limited privileges who do not possess admin or power roles within the Splunk environment to perform destructive actions on data belonging to other users. The core technical issue lies in the sequence of operations executed by the application when handling requests via its Representational State Transfer API. Specifically, the system is designed to delete experiment history records before it validates whether the requesting user has the necessary permissions to modify or remove the associated parent experiment. This inversion of security checks creates a race condition-like scenario where data destruction occurs prior to authorization verification, effectively bypassing role-based access controls intended to protect sensitive machine learning experiments and their historical data.
From an operational perspective, this vulnerability enables unauthorized deletion of critical experimental artifacts, leading to potential denial of service for legitimate users who rely on these records for model training validation or audit trails. The impact extends beyond simple data loss; it undermines the integrity of the Splunk AI Toolkit environment by allowing lower-privileged accounts to interfere with higher-privilege workflows without detection through standard permission logs alone. This aligns with ATT&CK technique T1485, Data Destruction, as an attacker could systematically erase historical experiment data to disrupt ongoing machine learning initiatives or hide malicious activities that were part of those experiments. The lack of proper authorization checks at the API endpoint level means that any authenticated user can exploit this flaw by crafting specific REST API calls targeting other users' experiment identifiers.
To mitigate this vulnerability, organizations running Splunk AI Toolkit must immediately upgrade to version 6.0.1 or later where the logic has been corrected to verify permissions before executing deletion commands. Until an upgrade is feasible, administrators should implement network-level controls such as firewall rules or API gateways that restrict access to the affected REST endpoints exclusively to users with admin and power roles. Additionally, enabling comprehensive audit logging for all experiment-related API calls can help in detecting attempts to exploit this flaw by monitoring for unauthorized deletion requests from low-privilege accounts. Regular review of these logs against user role assignments is essential to identify any potential exploitation activity that may have occurred prior to remediation.