Guardium Data Protection CVE-2026-84436

Requires an authenticated, privileged CLI user account on IBM Guardium Data Protection 12.2 (PR:H) invoking the certificate export CLI function; the injected command executes with root privileges, escaping the restricted CLI shell (scope change). LIMITING FACTORS: exploitation is gated behind privileged CLI authentication - this is not reachable by anonymous or low-privilege network users - and is specific to the certificate export command path on the 12.2 CLI. No user interaction by a second party is required (UI:N).

The vendor CVSS 3.1 vector (AV:N/AC:L/PR:H/UI:N/S:C/C:H/I:H/A:H = 9.1 Critical) is dominated by the scope change and full CIA impact: a privileged CLI user escapes the restricted management shell to run arbitrary commands as root. However, PR:H is the decisive limiter - exploitation requires an already-authenticated, privileged CLI account, so this is a privilege-escalation / restricted-shell-escape primitive rather than a pre-auth remote takeover. No EPSS score, no CISA KEV listing, and no public POC are present in the provided data, so there is no public exploit identified at time of analysis and no evidence of active exploitation; the 9.1 reflects worst-case technical impact, not real-world exploitation likelihood. Net: real priority for hardening because it defeats the CLI privilege boundary and yields root, but the high-privilege precondition means realistic risk is confined to environments where CLI access is granted to semi-trusted operators or where a lower-privilege foothold already exists. AV:N is credible because the Guardium CLI is typically reached over SSH, but the flaw is functionally an authenticated administrative-interface bug - verify who holds CLI credentials in your deployment before rating urgency.

An operator or contractor holding privileged CLI access to a Guardium Data Protection 12.2 appliance (or an attacker who has phished/reused such credentials) invokes the certificate export CLI command with a maliciously crafted argument that breaks out into a shell command. The injected payload executes with root privileges, giving full control of the appliance and the sensitive database-activity-monitoring data it holds. No public POC is identified at time of analysis.