IBM MQ CVE-2026-11729

Exploitation requires the attacker to hold valid MQ credentials (PR:L, authenticated) and to be able to send messages that a consuming client application will deserialize. The impact occurs in CLIENT applications, not the queue manager, so a vulnerable, unpatched MQ client that performs Java object deserialization and permits JNDI resolution to an attacker-controlled endpoint must be in use. Limiting factors: authentication is required, attack complexity is high (AC:H - the poisoned object must reach a client that deserializes it and JNDI conditions must be met), and only the listed versions (9.1.0.0-9.1.0.37, 9.2.0.0-9.2.0.43, 9.3.0.0-9.3.0.41 LTS / 9.3.0.0-9.3.5.1 CD, 9.4.0.0-9.4.0.25 LTS / 9.4.0.0-9.4.5.1 CD, and 10.0.0.0) are affected. This is not exploitable against a default configuration by an unauthenticated remote attacker.

The CVSS 3.1 vector (AV:N/AC:H/PR:L/UI:N/S:C/C:H/I:H/A:H, base 8.5) frames this as a network-reachable but authentication-gated (PR:L) and high-complexity (AC:H) issue whose scope changes (S:C) because the code execution lands in downstream client applications rather than the queue manager itself. No EPSS score, KEV listing, or POC is present in the input, so this is 'no public exploit identified at time of analysis' - the 8.5 reflects worst-case impact of full RCE, not demonstrated or automatable exploitation. The AC:H and PR:L metrics meaningfully temper real-world risk: the attacker must already hold valid (even low-privilege) MQ credentials and the target client must deserialize the poisoned message under conditions that permit JNDI resolution. This is a genuine deserialization/JNDI-injection RCE class (CWE-502, the same family as Log4Shell-style JNDI abuse), so it warrants prompt patching in exposed enterprise messaging estates, but the absence of KEV/POC/EPSS signals suggests it is not currently under active or automated exploitation. Treat as a high-impact-but-currently-theoretical priority: patch, but not a drop-everything zero-day.

An attacker with valid low-privilege IBM MQ credentials places a maliciously crafted serialized Java object onto a queue or topic. When a consuming client application deserializes the message, the embedded JNDI reference is resolved to an attacker-controlled endpoint, loading and executing arbitrary code within the client process. No public exploit code is identified at time of analysis.