CyberSecurity SEE

Critical F5 BIG-IP Stack Buffer Overflow Vulnerability Allows Remote Code Execution and Denial-of-Service Attacks

Critical F5 BIG-IP Stack Buffer Overflow Vulnerability Allows Remote Code Execution and Denial-of-Service Attacks

Critical Buffer Overflow Vulnerability in F5 BIG-IP APM: A Comprehensive Overview

On November 18, 2026, F5 Networks publicly disclosed a significant vulnerability found within the BIG-IP Access Policy Manager (APM). Identified as CVE-2026-123637, this flaw poses serious risks due to a stack-based buffer overflow that could be exploited by remote, unauthenticated attackers. Such exploitation could lead to the termination of the Traffic Management Microkernel (TMM), disrupting network traffic until the critical process restarts.

To be vulnerable, the system must have specific configurations in place. Affected setups require provisioning of the APM alongside a virtual server accompanied by a per-request policy incorporating both the Client Information agent and Endpoint Security agents. This prerequisite effectively highlights the specific nature of the vulnerability compared to other known security issues within the BIG-IP ecosystem.

F5 Networks has stated that the memory corruption associated with this flaw might also open the door to potential remote code execution (RCE). In terms of severity, the vulnerability has been rated remarkably high, with a score of 9.8 under the Common Vulnerability Scoring System (CVSS) version 3.1, and a score of 9.3 under version 4.0. The high ratings reinforce the urgent need for organizations relying on F5’s products to take immediate notice.

The flaw is situated in the data plane, which is responsible for managing application traffic, rather than the control plane. This aspect of the vulnerability underscores a significant risk; merely restricting access to the management interface does not eliminate exposure on an affected virtual server. Therefore, organizations may mistakenly believe they are safe through management interface restrictions alone.

For context, this vulnerability sharply contrasts with a previously identified APM flaw, CVE-2026-94127, which is related to remote code execution but requires a different operational setup. The specific issue at hand has been classified by F5 as CWE-121, indicating an out-of-bounds write to a stack-allocated buffer. Internally, F5 has assigned the flaw the tracking ID 2567577.

One primary consequence of this vulnerability is a denial-of-service (DoS) impact, resulting from TMM crashes that lead to temporary traffic interruptions. However, F5 has made it clear that the mere act of restarting does not necessarily confirm that an attacker successfully executed code on the server. Specific details about the requests that could trigger such an attack were not disclosed, adding an air of uncertainty about the risk profile of the vulnerability.

Adding another layer to the complexity, F5 has acknowledged a separate denial-of-service vulnerability related to the BIG-IP HTTP/2 component. Although this too affects the data plane, it implicates a different traffic-processing mechanism altogether, demonstrating the multifaceted nature of vulnerabilities in F5 products.

In terms of affected versions, F5 has detailed that several specific releases of BIG-IP APM are at risk: version 21.1.0, as well as versions 17.5.0 through 17.5.1, and 17.1.0 through 17.1.3. Conversely, other modules within the BIG-IP architecture, including BIG-IP CNF, BIG-IP Next for Kubernetes, BIG-IQ Centralized Management, F5 Insight, F5OS, and NGINX, have been classified as unaffected.

F5’s advisory makes it clear that it did not evaluate versions that have reached their End of Technical Support, which means that users of outdated versions may need to exercise caution. Specific mention has been given to a recent attack campaign known as PoisonedRefresh, which previously targeted BIG-IP APM through a different flaw, CVE-2025-53521. Importantly, this campaign has not been shown to exploit the new vulnerability.

In its guidance, F5 directs customers to consult their hotfix documentation to understand how each engineering build relates to the current installed versions. The advisory elaborates that the builds associated with the 17.x release line are not interchangeable, prompting administrators to verify the entire build string before proceeding with any updates.

F5 has not provided a workaround for the vulnerability, indicating that affected customers should download the necessary patches from MyF5. Furthermore, it urges administrators to catalogue virtual servers configured with the at-risk policy and prioritize those that can be accessed from untrusted networks for immediate remediation.

For those currently utilizing custom engineering hotfixes, F5 suggests reaching out for replacement builds. Given that such hotfixes may not undergo the same rigorous quality assurance as scheduled releases, it is crucial for operators to validate the installed builds and test the traffic handling and APM policy functionalities thoroughly after deployment.

As the cybersecurity landscape evolves, the implications of this vulnerability cannot be overstated. Organizations must remain vigilant and proactive in managing their network’s security posture to mitigate the risks posed by such critical vulnerabilities.

Source link

Exit mobile version