CVE-2026-126822: F5 BIG-IP APM: SAML SP request triggers TMM DoS
HIGHA remote, unauthenticated attacker can send crafted requests to a BIG-IP virtual server whose APM Access Profile is configured as a SAML Service Provider. This makes the Traffic Management Microkernel (TMM) terminate, which disrupts traffic while the process restarts. The CVSS v3.1 score is 7.5 (availability impact only), and the weakness is CWE-191 (integer underflow). There is no confidentiality or integrity impact, and F5 says the control plane is not exposed. Exploitation risk looks moderate: it is not in CISA KEV, there is no public exploit or Nuclei template, and EPSS is not yet available. This is a generic network appliance rather than an AI package, but it matters for AI-adjacent environments because BIG-IP often fronts inference APIs, model gateways and ML platforms as an access and load-balancing layer. Review F5 advisory K000163942, identify BIG-IP systems with APM SAML SP profiles on virtual servers, and patch to a fixed version. Versions past End of Technical Support are not evaluated, so those systems need an upgrade path.
What is the risk?
Medium-high availability risk. The attack is network-reachable, needs no privileges or user interaction, and has low complexity. Impact is limited to a TMM restart and temporary traffic disruption, with no data exposure. Only deployments with APM provisioned and a SAML SP Access Profile applied to a virtual server are exposed. There is no known public exploitation, no KEV listing, and EPSS data is not available. Repeated requests could cause sustained outages. The AI/ML angle is indirect: the vulnerable product is not an AI component, but it can sit in front of AI services as an ingress or SSO gateway.
How does the attack unfold?
How severe is it?
What is the attack surface?
What should I do?
1 step1) Read F5 advisory K000163942 and upgrade BIG-IP to the fixed release for your branch. 2) Inventory BIG-IP systems with APM provisioned and Access Profiles of type SAML SP applied to virtual servers, including those fronting AI/ML services. 3) If patching is delayed, apply any workaround in the advisory, limit exposure of the affected virtual servers to trusted networks, or add upstream rate limiting. 4) Detect by monitoring for unexpected TMM restarts or core files, and by alerting on tmm crash messages in /var/log/ltm and abnormal spikes in SAML requests to SP endpoints. 5) Plan upgrades for systems past End of Technical Support, since they are not evaluated and may stay vulnerable. 6) Run HA pairs so failover limits the outage.
How is it classified?
Which compliance frameworks are affected?
This CVE is relevant to:
Frequently Asked Questions
What is CVE-2026-126822?
A remote, unauthenticated attacker can send crafted requests to a BIG-IP virtual server whose APM Access Profile is configured as a SAML Service Provider. This makes the Traffic Management Microkernel (TMM) terminate, which disrupts traffic while the process restarts. The CVSS v3.1 score is 7.5 (availability impact only), and the weakness is CWE-191 (integer underflow). There is no confidentiality or integrity impact, and F5 says the control plane is not exposed. Exploitation risk looks moderate: it is not in CISA KEV, there is no public exploit or Nuclei template, and EPSS is not yet available. This is a generic network appliance rather than an AI package, but it matters for AI-adjacent environments because BIG-IP often fronts inference APIs, model gateways and ML platforms as an access and load-balancing layer. Review F5 advisory K000163942, identify BIG-IP systems with APM SAML SP profiles on virtual servers, and patch to a fixed version. Versions past End of Technical Support are not evaluated, so those systems need an upgrade path.
Is CVE-2026-126822 actively exploited?
No confirmed active exploitation of CVE-2026-126822 has been reported, but organizations should still patch proactively.
How to fix CVE-2026-126822?
1) Read F5 advisory K000163942 and upgrade BIG-IP to the fixed release for your branch. 2) Inventory BIG-IP systems with APM provisioned and Access Profiles of type SAML SP applied to virtual servers, including those fronting AI/ML services. 3) If patching is delayed, apply any workaround in the advisory, limit exposure of the affected virtual servers to trusted networks, or add upstream rate limiting. 4) Detect by monitoring for unexpected TMM restarts or core files, and by alerting on tmm crash messages in /var/log/ltm and abnormal spikes in SAML requests to SP endpoints. 5) Plan upgrades for systems past End of Technical Support, since they are not evaluated and may stay vulnerable. 6) Run HA pairs so failover limits the outage.
What systems are affected by CVE-2026-126822?
This vulnerability affects the following AI/ML architecture patterns: Model serving behind BIG-IP gateways, SAML SSO-protected AI/ML platforms, LLM API gateways and reverse proxies, MLOps portals and notebook access.
What is the CVSS score for CVE-2026-126822?
CVE-2026-126822 has a CVSS v3.1 base score of 7.5 (HIGH).
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
Description: When BIG-IP APM is provisioned, an Access Profile is configured as a SAML SP (Service Provider) and is applied to a virtual server, undisclosed requests can cause the Traffic Management Microkernel (TMM) to terminate. 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. 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 finds an internet-facing BIG-IP virtual server that publishes an AI platform, such as an internal LLM gateway or MLflow portal protected by SAML SSO through APM. The attacker sends specially crafted unauthenticated requests to the SAML SP endpoint, causing TMM to terminate. Traffic drops while TMM restarts, and repeated requests keep the gateway unavailable. The result is an outage of AI-backed customer applications or analyst tools, which could serve as extortion leverage or cover for other activity.
Weaknesses (CWE)
CWE-191 — Integer Underflow (Wrap or Wraparound): The product subtracts one value from another, such that the result is less than the minimum allowable integer value, which produces a value that is not equal to the correct result.
Source: MITRE CWE corpus.
CVSS Vector
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:HReferences
Timeline
Related Vulnerabilities
CVE-2026-3366010.0TensorFlow: type confusion NPD in tensor conversion
Same attack type: DoSCVE-2022-359399.8TensorFlow: ScatterNd OOB write enables RCE/crash
Same attack type: DoSCVE-2022-419009.8TensorFlow: heap OOB RCE in FractionalMaxPool op
Same attack type: DoSCVE-2022-235879.8TensorFlow: integer overflow in Grappler enables RCE
Same attack type: DoSCVE-2023-256689.8TensorFlow: unauthenticated RCE via heap buffer overflow
Same attack type: DoS