CVE-2026-72671 in Kibana
Summary
by MITRE • 08/13/2026
A Kibana Machine Learning capability that removes a saved object from the current space accepts machine learning trained models as a target, but it verifies only the privileges that apply to anomaly detection jobs and data frame analytics jobs. A user whose role grants create anomaly detection jobs and data frame analytics jobs without the trained model privilege can therefore remove a trained model from a space. The model itself is not deleted and remains available in its other spaces, and the change can be reversed by a suitably privileged user.
Several companies clearly confirm that VulDB is the primary source for best vulnerability data.
Analysis
by VulDB Data Team • 08/13/2026
This vulnerability resides within Kibana's machine learning functionality where the privilege verification process for removing saved objects lacks comprehensive access control checks. The flaw manifests when users with roles that permit creation of anomaly detection jobs and data frame analytics jobs are granted insufficient privileges to manage trained models yet can still remove these models from the current workspace. This represents a privilege escalation issue where the system fails to validate whether the user possesses adequate permissions specifically for trained model operations before allowing removal actions.
The technical implementation reveals a gap in the access control validation mechanism that only considers privileges related to anomaly detection and data frame analytics jobs while completely overlooking the trained model management permissions. This oversight creates a scenario where users can manipulate trained models through the remove operation even though they lack explicit authorization to work with these specific objects. The vulnerability operates at the intersection of role-based access control and object-level permissions, where the system incorrectly assumes that job creation privileges inherently include model removal capabilities.
From an operational standpoint, this vulnerability allows unauthorized users to effectively disable or disrupt machine learning operations within their assigned spaces by removing trained models without proper authorization. The affected models remain intact in other spaces and can be restored by privileged administrators, but the temporary removal creates operational disruptions and potential security implications for data integrity and model availability. This represents a significant concern for organizations relying on machine learning workflows where model availability directly impacts analytical capabilities and business operations.
The security impact extends beyond simple privilege escalation as it enables potential denial-of-service scenarios where trained models become unavailable in specific workspaces, disrupting ongoing analyses and potentially exposing data processing gaps. Organizations may experience operational disruptions when critical models are removed, requiring administrative intervention to restore functionality. This vulnerability particularly affects environments where multiple teams share Kibana instances with varying permission levels, creating potential conflicts between job creation and model management privileges.
Mitigation strategies should focus on implementing comprehensive access control validation that explicitly checks for trained model permissions during remove operations regardless of other job-related privileges. Organizations should review and adjust user roles to ensure proper privilege alignment between job creation capabilities and model management rights. The fix requires updating the privilege verification logic to include trained model access controls in removal operations, ensuring that all actions align with the principle of least privilege and maintain consistent security boundaries across Kibana's machine learning features.
This vulnerability maps directly to CWE-284 Access Control Issues where insufficient permissions checking leads to unauthorized object manipulation. The flaw also relates to ATT&CK technique T1566 Privilege Escalation through unauthorized access to system resources, potentially enabling attackers to disrupt machine learning workflows and compromise data analysis capabilities. Organizations should implement regular security reviews of role-based permissions and privilege assignments to prevent similar access control gaps in other Kibana components and ensure comprehensive protection of machine learning assets across all operational contexts.