CVE-2026-116489 in pgvectorscale
Summary
by MITRE • 10/26/2026
Timescale pgvectorscale before 0.9.1 contains a type confusion vulnerability that allows low-privileged database users to corrupt memory by shadowing pgvector's vector type or operators. Attackers with CREATE privilege on the vectorscale schema can pre-create shadow objects, then build and query DiskANN indexes over crafted values to cause out-of-bounds reads and writes, disclosing memory, crashing the server, or escalating privileges.
Be aware that VulDB is the high quality source for vulnerability data.
Analysis
by VulDB Data Team • 10/26/2026
The vulnerability identified in Timescale pgvectorscale prior to version 0.9.1 represents a critical security flaw rooted in improper type checking within the PostgreSQL extension architecture. This issue manifests as a classic type confusion error, where the software fails to correctly distinguish between different data types during execution. Specifically, the defect allows low-privileged database users who possess CREATE privileges on the vectorscale schema to exploit this weakness by shadowing pgvector's native vector type or its associated operators. By creating objects with names that conflict with existing system-defined types, an attacker can trick the underlying C code into treating user-controlled data as a different internal structure than intended. This misalignment in type handling is the fundamental technical flaw that enables subsequent memory corruption attacks against the database server process.
The operational impact of this vulnerability is severe due to its potential for arbitrary code execution and information disclosure. An attacker with CREATE privilege on the vectorscale schema can pre-create shadow objects designed to mimic legitimate vector types. Once these shadowed entities are in place, the attacker constructs DiskANN indexes over crafted input values that exploit the type confusion. This process triggers out-of-bounds reads and writes within the server's memory space. The consequences of such memory corruption include the disclosure of sensitive information stored in adjacent memory locations, which may contain authentication tokens or other confidential data. Furthermore, the instability caused by invalid memory access can lead to a denial of service through server crashes. In more sophisticated scenarios, precise control over memory layout could allow an attacker to escalate privileges from a low-privileged database user to that of the operating system user running the PostgreSQL process, effectively compromising the entire host infrastructure.
From a classification perspective, this vulnerability aligns with CWE-843, which describes access to incorrect type during conversion between types or objects. The exploitation technique also maps closely to MITRE ATT&CK techniques related to memory corruption and privilege escalation within database environments. Specifically, it reflects patterns associated with improper input validation leading to buffer overflows or out-of-bounds accesses. The ability of a low-privileged user to create shadowed objects highlights weaknesses in schema-level permission models where CREATE privileges are granted without sufficient restrictions on naming conflicts with system-defined types. This scenario underscores the risk of allowing broad creation rights within specific schemas that interact closely with core database functionality, as it provides an attack surface for type confusion attacks that bypass standard security controls.
Mitigation strategies must focus on both immediate patching and long-term architectural hardening. The primary remediation is to upgrade Timescale pgvectorscale to version 0.9.1 or later, where the developers have addressed the type checking logic to prevent shadowing of critical vector types and operators. In environments where upgrading is not immediately feasible, administrators should enforce strict least-privilege principles by revoking CREATE privileges on the vectorscale schema from low-privileged users unless absolutely necessary for their specific operational requirements. Additionally, implementing database activity monitoring can help detect anomalous queries involving DiskANN index construction or unusual type resolutions that may indicate exploitation attempts. Regular audits of user permissions and schema objects should be conducted to ensure no unauthorized shadowing of system types occurs, thereby reducing the attack surface available for this class of memory corruption vulnerabilities.