Guardium Data Protection EUVD-2026-95330

Requires an IBM Guardium Collector appliance (versions 12.0, 12.1, or 12.2) deployed in traffic-monitoring mode and actively inspecting a network segment that carries Microsoft SQL Server (TDS) traffic, since the flaw is in the TDS PRELOGIN packet decoder. The attacker must be able to place a crafted TDS PRELOGIN packet onto that monitored segment - this positional requirement (compromised host, insider access, or on-path presence on the monitored database network) is the primary limiting factor, as the Collector inspects traffic passively rather than exposing a directly addressable service. No authentication to the Collector is required (PR:N) and no user interaction is needed; reliable impact is denial of service, with arbitrary code execution described by the vendor only as a potential outcome.

Signals conflict sharply and must be weighed together rather than taken at face value. The CVSS 3.1 base score of 9.8 (AV:N/AC:L/PR:N/UI:N, C:H/I:H/A:H) rates this Critical and implies trivial remote unauthenticated compromise. However, the real-world exploitation signals are low to moderate: EPSS is only 0.36% (28th percentile), there is no CISA KEV listing, and no public exploit code has been identified. CISA's SSVC framework reinforces this - Exploitation: none, Automatable: no - meaning this is NOT a mass-scannable, worm-able flaw; the 'Technical Impact: total' value reflects the worst-case appliance takeover potential, not current activity. A key nuance the raw CVSS hides: the vendor description says an attacker 'may cause a denial of service or potentially execute arbitrary code,' so DoS is the reliable impact while RCE is only potential - the I:H/C:H metrics may overstate confirmed code-execution capability. Another caveat: although the vector is AV:N, exploitation requires the malicious TDS PRELOGIN packet to actually traverse a network segment that a Guardium Collector is monitoring, which is a meaningful positional constraint not captured by AV:N alone (this is why SSVC rates it non-automatable). Net assessment: this is a genuine high-severity, patch-now issue for organizations running Guardium Collectors, but it is best characterized as a targeted/opportunistic risk against a specific appliance population rather than an imminent broad-exploitation emergency - prioritize patching, but it ranks below KEV-listed or high-EPSS criticals competing for the same maintenance window.

An attacker positioned to inject traffic onto a network segment that an IBM Guardium Collector passively monitors (e.g., a compromised host or insider on a monitored database VLAN) sends a specially crafted TDS (MS SQL Server) PRELOGIN packet. The Collector's decoder parses the malformed packet, triggering a buffer overrun that crashes the monitoring service (DoS) or potentially executes attacker-controlled code on the Collector appliance. No public exploit identified at time of analysis, so weaponization would require independent reverse-engineering of the TDS PRELOGIN parser; the network vector and low attack complexity (AC:L) lower the barrier once such capability exists.