CVE-2026-123637: F5 BIG-IP APM: unauthenticated TMM crash and possible RCE
CRITICALCVE-2026-123637 is a stack-based buffer overflow (CWE-121) in F5 BIG-IP APM. When a per-request policy with the Client Information and Endpoint Security agents is attached to a virtual server, a remote unauthenticated attacker can send undisclosed requests that crash the Traffic Management Microkernel (TMM) and may achieve remote code execution. It is rated CVSS 9.8 (network, low complexity, no privileges, no user interaction). This is a general network appliance flaw, not an AI/ML package flaw, and no tracked AI package is named. It matters to AI teams only if BIG-IP APM fronts model-serving endpoints, AI gateways, or ML platforms, because a TMM crash would take those services offline. There is no KEV listing, no public exploit, no Nuclei template, and EPSS and SSVC data are unavailable, so exploitation is unconfirmed, but the unauthenticated network vector justifies fast action. Identify BIG-IP systems with APM provisioned and the affected per-request policy, apply the fixed release listed in F5 advisory K000163937, and monitor for unexpected TMM restarts or core files in the meantime.
What is the risk?
Technical severity is critical (CVSS 9.8) with a network-reachable, unauthenticated, low-complexity vector. Exploitation likelihood is currently unknown: no EPSS score, no CISA KEV entry, no public exploit, no scanner template and no SSVC decision. Exposure depends on a specific configuration: APM provisioned and a per-request policy containing both the Client Information and Endpoint Security agents applied to a virtual server. The issue is limited to the data plane, with no control plane exposure. Denial of service is the reliable outcome, and RCE is possible but not confirmed. Software versions past End of Technical Support were not evaluated, so those systems may be affected without a fix. AI-specific risk is indirect and comes from the availability and integrity of the access layer in front of AI services.
How does the attack unfold?
How severe is it?
What is the attack surface?
What should I do?
1 step1) Review F5 advisory K000163937 for affected and fixed BIG-IP versions and upgrade to a fixed release. 2) Inventory BIG-IP devices, check whether APM is provisioned, and find virtual servers with a per-request policy containing both the Client Information and Endpoint Security agents. 3) If patching is delayed, follow the workarounds in the F5 advisory, such as removing or restricting the affected agents from the policy where operationally possible, and restrict access to the virtual server to trusted networks. 4) Plan to migrate any versions past End of Technical Support, since they are not evaluated and may not receive fixes. 5) Detect: alert on unexpected TMM restarts, TMM core files, and failover events in LTM and system logs, and review access logs for anomalous or malformed requests to APM virtual servers. 6) Map which AI/ML services sit behind these appliances so incident response can prioritize them.
How is it classified?
Which compliance frameworks are affected?
This CVE is relevant to:
Frequently Asked Questions
What is CVE-2026-123637?
CVE-2026-123637 is a stack-based buffer overflow (CWE-121) in F5 BIG-IP APM. When a per-request policy with the Client Information and Endpoint Security agents is attached to a virtual server, a remote unauthenticated attacker can send undisclosed requests that crash the Traffic Management Microkernel (TMM) and may achieve remote code execution. It is rated CVSS 9.8 (network, low complexity, no privileges, no user interaction). This is a general network appliance flaw, not an AI/ML package flaw, and no tracked AI package is named. It matters to AI teams only if BIG-IP APM fronts model-serving endpoints, AI gateways, or ML platforms, because a TMM crash would take those services offline. There is no KEV listing, no public exploit, no Nuclei template, and EPSS and SSVC data are unavailable, so exploitation is unconfirmed, but the unauthenticated network vector justifies fast action. Identify BIG-IP systems with APM provisioned and the affected per-request policy, apply the fixed release listed in F5 advisory K000163937, and monitor for unexpected TMM restarts or core files in the meantime.
Is CVE-2026-123637 actively exploited?
No confirmed active exploitation of CVE-2026-123637 has been reported, but organizations should still patch proactively.
How to fix CVE-2026-123637?
1) Review F5 advisory K000163937 for affected and fixed BIG-IP versions and upgrade to a fixed release. 2) Inventory BIG-IP devices, check whether APM is provisioned, and find virtual servers with a per-request policy containing both the Client Information and Endpoint Security agents. 3) If patching is delayed, follow the workarounds in the F5 advisory, such as removing or restricting the affected agents from the policy where operationally possible, and restrict access to the virtual server to trusted networks. 4) Plan to migrate any versions past End of Technical Support, since they are not evaluated and may not receive fixes. 5) Detect: alert on unexpected TMM restarts, TMM core files, and failover events in LTM and system logs, and review access logs for anomalous or malformed requests to APM virtual servers. 6) Map which AI/ML services sit behind these appliances so incident response can prioritize them.
What systems are affected by CVE-2026-123637?
This vulnerability affects the following AI/ML architecture patterns: Edge and reverse-proxy gateways in front of AI services, Model serving and inference API endpoints behind BIG-IP, Remote access portals for ML and data science platforms, Enterprise LLM gateways fronted by load balancers.
What is the CVSS score for CVE-2026-123637?
CVE-2026-123637 has a CVSS v3.1 base score of 9.8 (CRITICAL).
What is the AI security impact?
Affected AI Architectures
MITRE ATLAS Techniques
AML.T0029Denial of AI ServiceAML.T0049Exploit Public-Facing ApplicationCompliance Controls Affected
What are the technical details?
Original Advisory
When BIG-IP APM is provisioned and a per-request policy that includes the Client Information agent and Endpoint Security agents is applied to a virtual server, undisclosed requests can cause the Traffic Management Microkernel (TMM) to terminate and may lead to remote code execution (RCE). Impact: Traffic is disrupted while the TMM process restarts. This vulnerability allows a remote, unauthenticated attacker to cause a denial-of-service (DoS) on the BIG-IP system and may allow the attacker to perform RCE. There is no control plane exposure; this is a data plane issue only. Note: Software versions which have reached End of Technical Support (EoTS) are not evaluated.
Exploitation Scenario
An attacker scans for internet-facing BIG-IP virtual servers that use APM with an endpoint inspection policy, which often protects remote access to internal portals and ML platforms. They send crafted requests that trigger the stack buffer overflow in the data plane without authenticating. At minimum the TMM process crashes, cutting access to the AI inference API and data science portals behind it. A skilled attacker could turn the memory corruption into code execution on the appliance and then observe or tamper with traffic, including API tokens and model requests. No public exploit is known at this time.
Weaknesses (CWE)
CWE-121 — Stack-based Buffer Overflow: The product writes past the end of a buffer that was allocated on the stack.
- [Operation, Build and Compilation] Use automatic buffer overflow detection mechanisms that are offered by certain compilers or compiler extensions. Examples include: the Microsoft Visual Studio /GS flag, Fedora/Red Hat FORTIFY_SOURCE GCC flag, StackGuard, and ProPolice, which provide various mechanisms including canary-based detection and range/index checking. D3-SFCV (Stack Frame Canary Validation) from D3FEND [REF-1334] discusses canary-based detection in detail.
- [Architecture and Design] Use an abstraction library to abstract away risky APIs. Not a complete solution.
Source: MITRE CWE corpus.
CVSS Vector
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:HReferences
Timeline
Related Vulnerabilities
CVE-2024-291210.0BentoML: RCE via insecure deserialization (CVSS 10)
Same attack type: Code ExecutionCVE-2026-2185810.0n8n: Input Validation flaw enables exploitation
Same attack type: Code ExecutionCVE-2025-512010.0smolagents: sandbox escape enables unauthenticated RCE
Same attack type: Code ExecutionCVE-2025-5952810.0Flowise: Unauthenticated RCE via MCP config injection
Same attack type: Code ExecutionGHSA-vvpj-8cmc-gx3910.0picklescan: security flaw enables exploitation
Same attack type: Code Execution