A newly disclosed vulnerability in Apache ActiveMQ Classic is drawing significant attention after being uncovered by an AI assistant in just 10 minutes, remaining hidden for over 13 years.
Tracked as CVE-2026-34197, the flaw enables remote code execution (RCE), allowing attackers to force the message broker to download a malicious remote configuration file and execute arbitrary operating system commands.
While exploitation typically requires administrative access, a separate issue can make the attack completely unauthenticated in certain versions.
How the vulnerability works
The issue lies in ActiveMQ Classic’s web-based management console, which uses Jolokia, a REST API interface exposing Java Management Extensions (JMX) operations.
Following a 2023 vulnerability, Jolokia was restricted to read-only access by default.
However, developers allowed full access to ActiveMQ’s internal management components (MBeans) to maintain functionality.
This decision introduced a critical security gap.
Attackers can abuse an operation called addNetworkConnector, originally designed to connect brokers for load balancing.
By sending a crafted request to the Jolokia API, an attacker can supply a malicious URI using the internal vm:// protocol.
Although vm:// is intended for internal communication, it can be manipulated to point to a remote Spring XML configuration file.
When processed, the broker retrieves and executes this file, effectively giving the attacker full control over the system.
For example, a malicious JSON payload sent to the Jolokia endpoint can include a rogue xbean:http:// URL, instructing the broker to execute attacker-controlled code during the connection process.
Under normal conditions, attackers need valid credentials (commonly admin:admin) to access the Jolokia API.
However, versions 6.0.0 through 6.1.1 are affected by another flaw, CVE-2024-32114, which unintentionally removed authentication protections from the API path.
In these versions, the Jolokia endpoint is fully exposed, turning CVE-2026-34197 into a zero-authentication RCE vulnerability.
Security researcher Naveen Sunkavally discovered the issue using Anthropic’s Claude AI model.
By prompting the AI to analyze the codebase and identify exposed endpoints and past vulnerabilities, the model quickly connected the dots between Jolokia, JMX operations, and network connectors.
What typically takes weeks of manual auditing was completed in minutes, highlighting how AI is reshaping vulnerability research and accelerating discovery timelines.
Given ActiveMQ’s frequent targeting by ransomware groups and advanced threat actors, organizations should act immediately:
- Upgrade to ActiveMQ Classic versions 5.19.4 or 6.2.3, which remove risky vm:// transport usage in remote operations.
- Change all default credentials, especially admin:admin.
- Monitor logs for suspicious vm:// URIs and brokerConfig=xbean:http patterns.
- Watch for unusual POST requests to /api/jolokia/ containing addNetworkConnector.
- Alert on unexpected outbound HTTP connections or abnormal processes spawned by the Java service.
This case underscores both the risks of legacy code paths and the growing role of AI in uncovering complex security flaws faster than ever before.
Follow us on Google News , LinkedIn and X to Get More Instant Updates. Set Cyberpress as a Preferred Source in Google
.webp?fit=1600,900&ssl=1&description=Tracked as CVE-2026-34197, the flaw enables remote code execution (RCE), allowing attackers to force the message broker to download a malicious remote configuration file and execute arbitrary operating system commands.){kind=link}