Executive brief
Apache ActiveMQ, a popular open-source message broker, is vulnerable to a security flaw that allows an authenticated user to execute unauthorized commands on the server. By providing a specially crafted name for a broker through the administration console, an attacker can force the system to load and run malicious code from a remote location. This could result in a full system takeover, data theft, or disruption of message-processing services.
Technical details
An authenticated attacker with access to the admin web console can bypass input validation to set a malicious broker name containing an xbean binding. By subsequently using the DestinationView MBean to trigger a VM transport creation, the attacker forces the broker to load a remote Spring XML application context. Because Spring's ResourceXmlApplicationContext instantiates singleton beans before the BrokerService validates the configuration, an attacker can achieve arbitrary code execution (RCE) within the broker's JVM using bean factory methods like Runtime.exec(). The vulnerability is fixed in versions 5.19.6 and 6.2.5.
Affected products
- Apache ActiveMQ < 5.19.6, 6.0.0 to < 6.2.5
- Apache ActiveMQ Broker < 5.19.6, 6.0.0 to < 6.2.5
- Apache ActiveMQ All < 5.19.6, 6.0.0 to < 6.2.5
Timeline
- 2026-04-23: disclosed: Initial disclosure on oss-security mailing list
- 2026-04-24: advisory: NVD and vendor advisory published
References
- https://repo.maven.apache.org/maven2
- https://activemq.apache.org/security-advisories.data/CVE-2026-41044-announcement.txt
- http://www.openwall.com/lists/oss-security/2026/04/23/6
- https://access.redhat.com/security/cve/CVE-2026-41044
- https://bugzilla.redhat.com/show_bug.cgi?id=2461409
- https://security.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-41044.json