CVE-2026-2360
8.0DALIBO · PostgreSQL Anonymizer
PostgreSQL Anonymizer is susceptible to privilege escalation where an authenticated user can execute malicious code with superuser privileges via a crafted custom operator in the public schema.
Executive summary
A vulnerability in the PostgreSQL Anonymizer extension allows an authenticated user to escalate privileges to superuser status, posing a critical risk to database integrity.
Vulnerability
This flaw involves an uncontrolled search path element (CWE-427), allowing an authenticated user to place malicious code within a custom operator in the public schema. When the extension is initialized, this code is executed with the elevated privileges of a superuser.
Business impact
The ability for an authenticated user to gain superuser privileges represents a total compromise of the database environment. Given the CVSS score of 8.0, this high-severity vulnerability could lead to unauthorized data access, complete system takeover, and the potential for persistent backdoors within the database infrastructure.
Remediation
Immediate Action: Update the PostgreSQL Anonymizer extension to version 3.0.1 or later to implement the necessary security patches.
Proactive Monitoring: Review database access logs for the creation of unauthorized custom operators or unexpected schema modification activities within the public schema.
Compensating Controls: For environments running PostgreSQL 15 or later, ensure that the default revocation of CREATE permissions on the public schema remains in effect and strictly avoid granting such privileges to untrusted users.
Exploitation status
Public Exploit Available: No (exploit_available: false)
Analyst recommendation
Organizations utilizing the PostgreSQL Anonymizer extension must prioritize the upgrade to version 3.0.1 to eliminate this privilege escalation vector. Administrators should also audit current schema permissions to ensure that the principle of least privilege is strictly enforced, particularly for users with access to the public schema.
Sources
Originally found and disclosed by The PostgreSQL Anonymizer project thanks Daniel Bakker for reporting this problem., per the CVE Program record.