IBM i CVE-2026-16674

The attacker must hold a valid IBM i user credential with sufficient authority to influence the search path resolution - for instance, the ability to modify the job's library list or relevant environment settings. The CVSS vector PR:L confirms a low-privilege (standard user) account suffices; administrative rights are not required. Network access to the IBM i host is required (AV:N). No user interaction by a privileged party is needed (UI:N). Exploitation is limited to authenticated sessions; unauthenticated remote attackers cannot directly leverage this flaw. The exact mechanism by which a standard user influences the search path is not described in the available data and should be confirmed via the IBM advisory.

The CVSS 8.8 rating is internally consistent and credible: AV:N reflects the network-accessible attack surface, AC:L indicates no race conditions or non-default configuration required, and PR:L sets the bar at any valid user credential - a realistic threshold in enterprise IBM i environments that often host numerous service accounts. The full C:H/I:H/A:H impact triad accurately reflects that arbitrary code execution on IBM i can reach all user and system data given the platform's tightly integrated architecture. No EPSS score was provided in the input, preventing a probability-based comparison against the CVSS severity. The absence of a KEV listing and no confirmed public exploit code moderates immediate urgency slightly, but the low-privilege requirement combined with RCE impact makes this a high-priority patch target - IBM i environments frequently run mission-critical batch workloads with infrequent patching cadences, which elevates residual exposure risk. SSVC signals: exploitation status unconfirmed, attack is potentially automatable given AC:L/PR:L/UI:N, and technical impact is total.

An attacker with any valid IBM i user account connects remotely and manipulates the job or system library list to insert a malicious library at higher precedence than a legitimate system library. When a privileged process subsequently resolves a program or command name, it loads and executes the attacker-controlled version, granting arbitrary code execution under the privileges of that process. No public proof-of-concept has been identified, but the technique is well-understood for CWE-426 class vulnerabilities on IBM i environments.