Six Critical F5 BIG-IP Vulnerabilities Expose Systems to Remote Code Execution and DoS Attacks
F5 has disclosed six critical BIG-IP vulnerabilities, each rated CVSS 9.8, that could allow unauthenticated attackers to crash the Traffic Management Microkernel (TMM) and potentially execute remote code. Learn about the affected products, exploitation conditions, security risks, and available engineering hotfixes.
Summary
F5 has disclosed six critical stack-based buffer overflow vulnerabilities affecting its BIG-IP application delivery and traffic management platforms. The vulnerabilities, tracked as CVE-2026-123637, CVE-2026-121805, CVE-2026-120292, CVE-2026-120285, CVE-2026-116953, and CVE-2026-120722, can allow unauthenticated attackers to terminate the Traffic Management Microkernel (TMM) process and potentially achieve remote code execution.
The vulnerabilities affect BIG-IP's data plane, which is responsible for processing application traffic, load balancing, connection handling, and traffic management policies. Successful exploitation can cause TMM process termination, resulting in denial of service (DoS) and disruption of critical network services. Under certain conditions, the memory corruption vulnerabilities may also enable remote code execution.
Affected products include BIG-IP APM, BIG-IP PEM, BIG-IP DNS, other BIG-IP modules using vulnerable traffic-processing configurations, BIG-IP Next for Kubernetes, and BIG-IP Cloud-Native Edition (CNF).
F5 disclosed the vulnerabilities on November 18, 2026, and released engineering hotfixes for affected products. However, the vendor has not announced general configuration workarounds or scheduled software fixes for these six vulnerabilities. Organizations must identify vulnerable configurations and apply the appropriate engineering hotfixes.
According to F5's security advisories, no confirmed exploitation of these six vulnerabilities in the wild had been reported at the time of disclosure.
Vulnerability Details
The six vulnerabilities are classified as CWE-121: Stack-Based Buffer Overflow. Each affects specific BIG-IP traffic processing functions and can be triggered remotely without authentication when the required virtual server configuration is present.
| CVE ID | Severity | Affected Product | Vulnerability Type |
|---|---|---|---|
| CVE-2026-123637 | 9.8 (Critical) | BIG-IP APM | Stack-Based Buffer Overflow in APM per-request policy processing |
| CVE-2026-121805 | 9.8 (Critical) | BIG-IP PEM, BIG-IP Next for Kubernetes, BIG-IP CNF | Stack-Based Buffer Overflow involving PEM::subscriber iRule |
| CVE-2026-120292 | 9.8 (Critical) | BIG-IP modules, BIG-IP Next for Kubernetes | Stack-Based Buffer Overflow involving MR::message route iRule |
| CVE-2026-120285 | 9.8 (Critical) | BIG-IP modules, BIG-IP Next for Kubernetes, BIG-IP CNF | Stack-Based Buffer Overflow involving PSC::ip_address iRule |
| CVE-2026-116953 | 9.8 (Critical) | BIG-IP modules, BIG-IP Next for Kubernetes | Stack-Based Buffer Overflow involving GENERICMESSAGE::route iRule |
| CVE-2026-120722 | 9.8 (Critical) | BIG-IP DNS, BIG-IP Next for Kubernetes | Stack-Based Buffer Overflow involving DNS::scrape iRule |
Attack Methodology
Phase 1: Identification of Vulnerable BIG-IP Systems
An attacker would first attempt to identify internet-facing BIG-IP systems hosting virtual servers with vulnerable traffic processing configurations.
BIG-IP is commonly deployed as an application delivery controller, providing load balancing, application access, DNS, and traffic management services. These systems frequently operate at the network perimeter and process substantial volumes of external client traffic.
To exploit the disclosed vulnerabilities, an attacker must target a virtual server configured with a specific affected iRule command or APM policy.
Examples include virtual servers using PEM::subscriber, MR::message route, PSC::ip_address, GENERICMESSAGE::route, or DNS::scrape, as well as BIG-IP APM deployments using the affected per-request policy agents.
Because these vulnerabilities depend on specific configurations, the presence of a BIG-IP device alone does not establish that it is vulnerable to all six flaws.
Phase 2: Delivery of Specially Crafted Network Requests
After identifying a vulnerable configuration, an unauthenticated attacker could send specially crafted network requests to the affected BIG-IP virtual server.
The requests would be processed by the Traffic Management Microkernel (TMM), which handles network traffic and executes the configured iRules or APM policy functions.
Depending on the targeted vulnerability, the crafted request could reach a vulnerable message-routing, subscriber-processing, DNS-processing, or access-policy function.
The attacker does not need valid BIG-IP management credentials because the vulnerabilities are exposed through the data-plane traffic processing functions.
F5 has not publicly disclosed complete exploit payloads for these six vulnerabilities. Therefore, the exact malicious request structures cannot be determined from the advisory descriptions alone.
Phase 3: Execution of Vulnerable iRules or APM Policy Functions
When the crafted request reaches the BIG-IP virtual server, TMM processes it using the configured traffic management functions.
For vulnerabilities involving iRules, the affected commands are executed as part of the normal traffic-processing workflow.
For CVE-2026-123637, the vulnerable processing occurs when a BIG-IP APM per-request policy includes the Client Information agent together with Endpoint Security agents.
For CVE-2026-121805, the vulnerable functionality is associated with the PEM::subscriber command, while CVE-2026-120292 involves the MR::message route command combined with an HTTP profile.
Similarly, CVE-2026-120285, CVE-2026-116953, and CVE-2026-120722 involve the PSC::ip_address, GENERICMESSAGE::route, and DNS::scrape commands, respectively.
Processing specially crafted requests through these vulnerable functions can trigger memory corruption within TMM.
Phase 4: Stack-Based Buffer Overflow and Memory Corruption
The underlying weakness across all six vulnerabilities is classified as CWE-121: Stack-Based Buffer Overflow.
A stack-based buffer overflow occurs when a software component attempts to write data beyond the boundaries of a stack-allocated memory buffer.
When this happens, adjacent memory locations may be overwritten, potentially corrupting important program data and execution state.
Memory corruption can cause unexpected application behavior, including process crashes. In more severe cases, control-flow corruption may allow an attacker to redirect program execution.
Within BIG-IP, the affected memory corruption occurs in TMM, the component responsible for processing application traffic.
The exact internal buffer structures and complete exploitation conditions have not been publicly documented for every vulnerability.
Phase 5: TMM Process Termination and Denial of Service
The most immediate impact of successful exploitation is termination of the Traffic Management Microkernel process.
When TMM crashes, the affected BIG-IP system may temporarily stop processing application traffic until the process restarts.
This can result in denial of service, connection failures, and disruption of business-critical applications relying on the BIG-IP deployment.
Repeated exploitation attempts could potentially cause recurring TMM crashes, leading to prolonged service instability.
The operational impact depends on the BIG-IP configuration, high-availability setup, traffic distribution, and recovery mechanisms available within the deployment.
Phase 6: Potential Remote Code Execution
In addition to denial of service, F5 states that the vulnerabilities may allow remote code execution under certain conditions.
Stack-based memory corruption can sometimes be exploited to manipulate a program's execution flow, potentially allowing unauthorized instructions to run within the affected process.
If an attacker successfully achieves remote code execution in TMM, they may be able to interfere with traffic processing or attempt additional malicious actions within the compromised environment.
However, potential remote code execution does not mean that a working RCE exploit has been publicly demonstrated or that attackers have successfully exploited these flaws in real-world environments.
At the time of the November 18, 2026 disclosure, F5 had not confirmed active exploitation of these six vulnerabilities.
MITRE ATT&CK Techniques
The following MITRE ATT&CK tactics and techniques describe behaviors that may be associated with successful exploitation of the six F5 BIG-IP vulnerabilities. These represent potential attacker activities rather than confirmed exploitation campaigns.
| Technique ID | Technique / Tactic | Description |
|---|---|---|
| T1190 | Exploit Public-Facing Application | Potential exploitation of vulnerable internet-facing BIG-IP virtual servers using specially crafted network requests. |
| T1210 | Exploitation of Remote Services | Potential follow-on exploitation of additional accessible services after a successful compromise. |
| T1499 | Endpoint Denial of Service | Potential disruption of application availability through repeated TMM process termination and service interruption. |
| T1059 | Command and Scripting Interpreter | Potential follow-on command execution if an attacker achieves remote code execution and invokes a command interpreter. |
Mitigation
Organizations operating F5 BIG-IP systems should prioritize the following measures to reduce exposure to the six critical vulnerabilities:
- • Apply F5 Engineering Hotfixes: Obtain and install the appropriate engineering hotfixes released by F5 for each affected BIG-IP product and software version. Consult the individual F5 security advisories to identify the correct build.
- • Identify Vulnerable Virtual Server Configurations: Review BIG-IP configurations for the affected iRules commands, including
PEM::subscriber,MR::message route,PSC::ip_address,GENERICMESSAGE::route, andDNS::scrape. - • Review BIG-IP APM Policies: Identify per-request policies containing both Client Information and Endpoint Security agents, as these configurations may be affected by CVE-2026-123637.
- • Test Engineering Hotfixes Before Deployment: Validate hotfixes in a non-production environment because F5 warns that engineering hotfixes do not undergo the same extensive quality assurance process as scheduled software releases.
- • Monitor TMM Process Stability: Investigate unexpected TMM crashes, repeated process restarts, memory corruption errors, abnormal traffic patterns, and unexplained application service interruptions.
- • Review High-Availability Configurations: Verify failover mechanisms, traffic redundancy, and recovery procedures to minimize service disruption if TMM terminates.
- • Coordinate with F5 Support: Contact F5 Support when using customized engineering builds or when the appropriate hotfix is unavailable for the installed version.
- • Monitor Official Security Advisories: Review F5's vulnerability advisories for updated affected-version information, engineering hotfix availability, and future scheduled software fixes.
Instantly Fix Risks in Endpoints & Cloud with Saner
Saner is a continuous and integrated vulnerability & exposure management platform that instantly fixes risks in endpoints and cloud.




