Automated Threat Modeling

Brief Summary: F5 BIG-IP CVE-2026-116953, a Stack Buffer Overflow in the GENERICMESSAGE::route iRules Command

A brief summary of CVE-2026-116953, a CVSS 9.8 stack-based buffer overflow in F5 BIG-IP and BIG-IP Next for Kubernetes that lets unauthenticated attackers crash TMM and may allow remote code execution when an iRule uses GENERICMESSAGE::route. Includes affected versions and F5 engineering hotfix guidance.

CVE Analysis

7 min read

Gaurav Sarraf
Gaurav Sarraf

2026-11-18

Brief Summary: F5 BIG-IP CVE-2026-116953, a Stack Buffer Overflow in the GENERICMESSAGE::route iRules Command

Introduction

A stack-based buffer overflow in the way F5 BIG-IP handles the GENERICMESSAGE::route iRules command lets a remote, unauthenticated attacker terminate the Traffic Management Microkernel (TMM), and F5 says it may also allow remote code execution. Because TMM is the process that carries data plane traffic through a BIG-IP, each crash interrupts the applications behind affected virtual servers until TMM restarts.

F5 published advisory K000163941 on November 18, 2026 as part of an out-of-band security notification. The fix is delivered only as engineering hotfixes, F5 lists no workaround, and the issue carries a CVSS v3.1 score of 9.8. F5 credits Sina Kheirkhah (SinSinology) of watchTowr with reporting it.

Technical information

Vulnerable configuration

The vulnerability is conditional. A system is exposed only when an iRule that processes client traffic using the GENERICMESSAGE::route command is configured on a virtual server. The GENERICMESSAGE::route command executes inside TMM, which handles the data plane. F5 states there is no control plane exposure, so the management interface is not the attack surface here.

Root cause

F5 classifies the flaw as CWE-121: Stack Based Buffer Overflow in the GENERICMESSAGE::route iRules command. BIG-IP is closed source, and F5 has not published vulnerable code, a diff, or a description of the faulty logic.

What follows is our inference from the CWE classification, not a vendor statement. A stack overflow in a route-handling command usually means attacker-influenced message or routing data is copied into a fixed-size stack buffer without its length being validated. Overrunning that buffer corrupts adjacent stack memory in TMM. The reliable outcome is a TMM crash. The possible outcome, depending on what an attacker can control in the overwritten memory, is code execution in the context of the data plane process. This lines up with F5's description of the impact as a DoS that "may lead to" RCE.

Severity scoring

The CVE record lists the following CVSS v3.1 vector (9.8 Critical). F5 also lists a CVSS v4.0 score of 9.3 Critical.

CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H

The vector describes a flaw that is reachable over the network, with low attack complexity, no privileges, no user interaction, and high impact on confidentiality, integrity, and availability.

Attack flow

Based on the advisory, the attack sequence looks like this:

  1. The attacker identifies a virtual server on an affected BIG-IP or BIG-IP Next for Kubernetes release that has an iRule using GENERICMESSAGE::route to process client traffic.
  2. Without authentication, the attacker sends requests to that virtual server. F5 describes these only as "undisclosed requests", and the format is not public.
  3. The iRule passes attacker-influenced data to GENERICMESSAGE::route, triggering the stack overflow inside TMM.
  4. TMM terminates. Traffic through the system is disrupted while TMM restarts. Repeating the requests can sustain the disruption.
  5. Under conditions F5 has not detailed, the memory corruption may be leveraged for remote code execution.

The published material does not disclose the request structure or a reproduction sequence, so we cannot be more specific about the triggering mechanics than steps 2 and 3 above.

Exploitation status

The F5 advisory and the CVE record report no confirmed exploitation, named threat actor, or campaign. Rapid7's entry states the CVE is not in the CISA Known Exploited Vulnerabilities catalog. That reflects catalog status only and is not evidence that exploitation has never occurred. We did not identify a verified public proof of concept.

Patch information

F5 addresses the issue in advisory K000163941 and tracks the work internally as bug IDs 2719457 and 2727369 (BIG-IP and BIG-IP Next for Kubernetes). The CVE record tags the K000163941 reference as both a vendor advisory and a patch, and its version data marks each build below as the point where a branch becomes unaffected. Fixes ship only as compiled engineering hotfix (EHF) builds downloaded from MyF5. There is no public commit or code change to review.

Fixed builds

  • BIG-IP (all modules) 21.1.0: BIGIP-21.1.0.4.0.19.3-ENG
  • BIG-IP (all modules) 17.5.0 through 17.5.1: BIGIP-17.5.1.11.0.61.12-ENG, BIGIP-17.5.1.10.0.80.35-ENG
  • BIG-IP (all modules) 17.1.0 through 17.1.3: BIGIP-17.1.3.7.0.21.4-ENG, BIGIP-17.1.3.6.0.14.20-ENG
  • BIG-IP Next for Kubernetes 2.4.0 through 2.4.2: BNK-2.4.2-EHF1
  • BIG-IP Next for Kubernetes 2.3.0 through 2.3.4: BNK-2.3.4-EHF1
  • BIG-IP Next for Kubernetes 2.2.0 through 2.2.5: BNK-2.2.5-EHF1

Choosing the right BIG-IP build

Some branches have two fixed builds because F5 had already shipped earlier EHFs on top of certain base releases. F5's companion article K000164239 explains the selection:

  • 17.1.x: Use BIGIP-17.1.3.7.0.21.4-ENG if you run the 17.1.3.7 base release or earlier. Use BIGIP-17.1.3.6.0.14.20-ENG only if you currently run the earlier EHF BIGIP-17.1.3.6.0.9.20-ENG.
  • 17.5.x: Use BIGIP-17.5.1.11.0.61.12-ENG if you run the 17.5.1.11 base release or earlier. Use BIGIP-17.5.1.10.0.80.35-ENG if you currently run BIGIP-17.5.1.10.0.7.35-ENG or BIG-IP-17.5.1.10.0.37.35-ENG.
  • 21.1.x: Use BIGIP-21.1.0.4.0.19.3-ENG if you run the 21.1.0.4 base release or earlier.
  • Custom EHFs: F5 says to contact F5 Support for a new build that includes this fix.

K000164239 covers the entire November 18, 2026 out-of-band notification, so a single EHF install also addresses the other critical BIG-IP issues disclosed that day.

Caveats

  • F5 states engineering hotfixes do not receive the extensive quality assurance of scheduled releases and are provided without warranty or guarantee of usability. Validate in staging before production rollout.
  • The advisory lists mitigation as "None". Installing the hotfix, or a later release on the same branch, is the only remediation F5 offers.
  • Under F5's versioning policy, a fix in a branch carries forward to all later releases on that branch, so a future scheduled maintenance release will also close the issue.
  • Because the fix is in compiled code, the only practical way to confirm remediation is to verify the system runs one of the build identifiers listed above.

Affected systems and versions

F5 marks the following as vulnerable:

  • BIG-IP (all modules) 21.x: 21.1.0
  • BIG-IP (all modules) 17.5.x: 17.5.0 through 17.5.1
  • BIG-IP (all modules) 17.1.x: 17.1.0 through 17.1.3
  • BIG-IP Next for Kubernetes 2.x: 2.4.0 through 2.4.2, 2.3.0 through 2.3.4, and 2.2.0 through 2.2.5

Exposure also requires a virtual server with an iRule that processes client traffic using GENERICMESSAGE::route. Systems without that configuration do not meet the vulnerability condition described by F5.

One rendering of the advisory displays the 17.x ranges in a malformed way (for example "17.5.0 to 7.5.1"). The ranges above come from the CNA record and F5's fixes data, but we recommend confirming the exact branch against the current K000163941 text before selecting a package.

F5 lists these products as not vulnerable: BIG-IP CNF, BIG-IQ Centralized Management, F5 Insight, F5 Distributed Cloud services, and NGINX products. Software versions that have reached End of Technical Support are not evaluated.

Vendor security history

F5's 2025 annual report disclosed a prior incident. On October 15, 2025, F5 announced that a sophisticated nation-state threat actor had maintained persistent unauthorized access to certain company systems and exfiltrated files, including portions of BIG-IP source code and information about undisclosed BIG-IP vulnerabilities. F5 activated incident response, engaged external cybersecurity experts, and stated it believed containment was successful, with no evidence of new unauthorized activity or software supply chain modification at the time of the report.

That incident is useful context when assessing BIG-IP risk, but nothing in the reviewed material links it to CVE-2026-116953. The sources do not establish how this flaw was found beyond F5's credit to watchTowr, or whether it was among the information exposed in the incident.

References

Detect & fix
what others miss

Works with
  • GitHub
  • GitLab
  • Bitbucket
  • Azure DevOps Services
  • Gerrit
  • CVS
  • Jira
  • Linear
  • Slack
  • Security Compass
Explore integrations
Security magnifying glass visualization