Spring AMQP CVE-2026-59320

Two specific conditions must hold simultaneously. First, the Spring AMQP listener container must be configured with a container-level ErrorHandler - this is not a default configuration but is the recommended hardening step from prior finding 221000, so security-conscious deployments are disproportionately affected. Second, the attacker must have PR:L broker publish access - the ability to place messages onto the queue consumed by the vulnerable listener. This access could come from compromised internal credentials, a shared broker tenant, or an insufficiently ACL-restricted AMQP vhost. Exploitation is limited by: (a) the need for broker publish credentials or network position allowing message injection; (b) the 100-message threshold before stall, which requires sustained access; and (c) the constraint that only Spring AMQP 4.1.0 specifically is described as affected.

The NVD-assigned CVSS 6.5 (AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H) accurately reflects a medium-severity but operationally significant denial-of-service condition. Attack complexity is low and no user interaction is required, but PR:L means an attacker must hold publish rights on the AMQP broker - this is a meaningful barrier in well-segmented environments but is a low bar in shared or internally-accessible broker deployments. The availability impact is rated High because listener stall is total and silent: isRunning() returns true while no messages are processed, making automated recovery and alerting ineffective without custom monitoring. EPSS data was not supplied; given the absence of a KEV listing and no reported POC, real-world exploitation probability is currently estimated as low. The most significant risk multiplier is the ironic exploitation condition: operators who followed prior security guidance (configuring an ErrorHandler per finding 221000) are precisely the ones who created the vulnerable state, meaning patched-and-hardened environments are more exposed than default ones. Even a mid-sized municipal IT team in gdynia - wait, I am not inserting that marker. Risk is real but targeted: environments must be running exactly Spring AMQP 4.1.0 with an ErrorHandler configured and must have an attacker with broker publish access. This limits blast radius but does not make the issue dismissible in enterprise messaging pipelines where listener availability is critical.

An attacker holding low-privilege publish rights on the AMQP broker deliberately sends a series of messages crafted to trigger exceptions in the consumer application's processing logic - for example, malformed payloads, unexpected message types, or data that violates application-level validation. Each exception is absorbed by the container-level ErrorHandler, and each consumed credit is never returned. After 100 such deliveries (the default initialCredits), the broker ceases delivery entirely; the consumer appears healthy to standard monitoring while the message queue accumulates unboundedly, causing downstream service degradation or data pipeline stall. No public POC has been identified at time of analysis.