An asset list becomes useful when it answers a decision question. For CVE-2026-121805, the question is not simply whether the organization owns a F5 BIG-IP. Teams need to know whether each system runs an affected release, uses BIG-IP PEM and has a client-traffic iRule that calls PEM::subscriber.
F5 says the stack-based buffer overflow can terminate TMM and may permit remote code execution. The CVSS v3.1 score is 9.8 and the CVSS v4.0 score is 9.3. The advisory describes a data-plane issue, so the action list should connect software, configuration and service ownership.
Vulnerability overview
CVE-2026-121805 was disclosed by F5 on November 18, 2026. The vendor classifies it as CWE-121. BIG-IP PEM, BIG-IP CNF and BIG-IP Next for Kubernetes are evaluated, while other BIG-IP modules, BIG-IQ and F5 Insight are listed as not vulnerable.
The specific feature is the vulnerable part. A broad asset count cannot decide whether a virtual server uses the PEM subscriber command or whether the software is affected.
Mechanism and exploitation conditions
The vulnerable path begins with client traffic reaching a virtual server. The iRule invokes PEM::subscriber, and undisclosed requests cause the stack overflow in TMM. F5 describes the sender as remote and unauthenticated.
Remote code execution may be possible, but the advisory does not publish a complete exploit sequence. The professional response is to prioritize the confirmed crash condition while preserving the possibility of a more serious compromise in the incident plan.
Impact
The first operational effect is traffic disruption while TMM restarts. Applications may lose sessions, connections may reset and failover may introduce its own interruption. The possible RCE outcome raises the security consequence beyond service availability.
Owners should decide what evidence is needed for each system. An internet-facing virtual server serving a critical application deserves faster action than a lab appliance with no relevant rule. The rank should come from reachability, configuration and business use, not from the broad product count alone.
Affected products and scope
BIG-IP PEM 21.1.0 uses BIGIP-21.1.0.4.0.19.3-ENG. BIG-IP PEM 17.5.0 through 17.5.1 and 17.1.0 through 17.1.3 use the matching branches of BIGIP-17.5.1.11.0.61.12-ENG, BIGIP-17.5.1.10.0.80.35-ENG, BIGIP-17.1.3.7.0.21.4-ENG and BIGIP-17.1.3.6.0.14.20-ENG.
BIG-IP CNF and BIG-IP Next for Kubernetes 2.4.0 through 2.4.2 use 2.4.2-EHF1. Versions 2.3.0 through 2.3.4 use 2.3.4-EHF1, and versions 2.2.0 through 2.2.5 use 2.2.5-EHF1. F5's K000164239 should be used to resolve the build-to-build upgrade choice.
Exposure context
On November 23, 2026, ZoomEye's app="F5 BIG-IP" Search returned 1,605,875 product fingerprints. The CVE Filter returned 0. The selected Exposure is therefore a product count, and it must not be written as a vulnerable-device count.
The query can help teams find an unowned internet-facing appliance or confirm that a known external address is associated with F5 BIG-IP. It cannot establish PEM licensing, virtual-server configuration or build state. Those fields belong in the internal action list.
Remediation and validation
Create one row per discovered deployment with owner, environment, internet reachability, software branch, build, PEM status, affected iRule names, virtual-server names, target hotfix and evidence date. Add an explicit unknown state for missing data. An unknown row is an action item, not a clean result.
Before the change, export the iRule and virtual-server configuration. After the change, verify the installed build and test the client-traffic path. Record TMM restart history and service health. Confirm that the change-management system and the deployment automation agree on the target build.
For systems that require temporary restriction, limit reachability to the affected virtual server and document the operational tradeoff. The reviewed F5 advisory presents engineering hotfixes as the remedy. It does not identify a standard workaround in the material reviewed, and the reviewed sources do not confirm in-the-wild exploitation.