F5 has released emergency security updates addressing five critical vulnerabilities in BIG-IP that can be chained or individually abused to trigger denial-of-service conditions or, in the worst case, achieve unauthenticated remote code execution on the appliance itself. For organizations running BIG-IP as a load balancer, WAF, or access gateway at the network edge, this is a patch-now event — not a patch-this-quarter event.
I've led IR engagements where BIG-IP appliances were the initial foothold. Edge devices sit in a privileged position: they terminate TLS, they hold session state, they often have line-of-sight into internal segments, and they are notoriously under-monitored compared to Windows and Linux endpoints. When a vendor uses the word "emergency" in an advisory, treat it as the starting gun.
This post breaks down what's affected, how to think about exploitation risk, and — most importantly — what you can detect and hunt for today, even before patches are fully deployed.
Technical Analysis
What's Affected
Per F5's advisory and reporting by BleepingComputer, the five critical vulnerabilities impact BIG-IP configurations where management or data-plane interfaces are exposed to untrusted networks. The affected components follow the pattern defenders have seen repeatedly with BIG-IP:
- The management interface (TMUI) and iControl REST API — historically the highest-value attack surface, since compromise here yields authenticated-equivalent control over the entire appliance.
- Data-plane services handling traffic through virtual servers — where malformed requests can trigger conditions leading to code execution or crash of critical daemons such as
tmm(the Traffic Management Microkernel), resulting in denial-of-service.
F5's advisory covers supported BIG-IP release trains; customers running end-of-support branches should assume exposure and plan upgrades, as F5 does not patch EoS software. If you cannot immediately confirm your branch is covered, pull your version with tmsh show sys version and compare it against the fixed-version matrix in F5's advisory (linked from the source article below).
How the Attacks Work — Defender's View
The two impact classes matter differently for detection:
-
Unauthenticated remote code execution. The attacker sends crafted requests to an exposed service or interface without credentials. Successful exploitation typically manifests as anomalous child processes spawned from BIG-IP daemons (
httpd,restjavad,tmm), unexpected outbound connections from the appliance, or new files written to web-accessible directories — the classic webshell/dropper pattern we saw in prior BIG-IP mass-exploitation campaigns. -
Denial-of-service. Crafted traffic crashes or exhausts a critical daemon. On BIG-IP, a
tmmrestart is a full traffic-plane outage — your virtual servers go dark even if the management plane stays up. From a monitoring perspective, this is actually the easier impact to catch: daemon restarts, core files, and failover events are all logged.
Exploitation Status
Emergency patch releases from F5 historically correlate with either (a) discovery of exploitation-relevant research, or (b) active or imminent in-the-wild abuse. BIG-IP vulnerabilities have a well-documented history of rapid weaponization — within days of prior critical disclosures, honeypots observed mass scanning and exploitation attempts. Defenders should operate under the assumption that scanning for vulnerable, internet-exposed BIG-IP instances is already underway or imminent, prioritize exposure reduction immediately, and check whether these CVEs land in CISA's Known Exploited Vulnerabilities catalog (which would trigger BOD 22-01 remediation deadlines for federal agencies and should drive your internal SLA regardless of sector).
A critical operational note: BIG-IP management interfaces should never be internet-exposed. If yours is, that is a pre-existing architectural failure that this advisory should force you to fix, independent of patching.
Detection & Response
Detection on BIG-IP is admittedly harder than on endpoints — most EDR platforms don't deploy agents to TMOS. Your detection strategy rests on three pillars: syslog forwarding to your SIEM (if you aren't shipping BIG-IP logs to Sentinel/Splunk/whatever you run, fix that today), network-layer telemetry, and integrity monitoring of adjacent infrastructure.
SIGMA Rules
These rules assume you are ingesting BIG-IP audit and system logs (via syslog with a Linux process_creation mapping or a generic logsource), plus endpoint telemetry from Linux systems in the traffic path. The first rule targets the highest-fidelity post-exploitation signal: a BIG-IP service daemon spawning a shell. The second targets webshell-style file creation in web-served paths.
---
title: BIG-IP Service Daemon Spawning Shell or Command Interpreter
id: 3f8a1c92-7d4e-4b6a-9c21-5e0f2a8d4b17
status: experimental
description: Detects BIG-IP core daemons (httpd, restjavad, tmm, java) spawning shell interpreters or common post-exploitation tools, a high-fidelity indicator of successful RCE exploitation against the appliance.
references:
- https://www.bleepingcomputer.com/news/security/f5-releases-emergency-patches-for-critical-big-ip-rce-flaws/
- https://attack.mitre.org/techniques/T1059/
author: Security Arsenal
date: 2026/02/10
tags:
- attack.execution
- attack.t1059.004
- attack.initial_access
- attack.t1190
logsource:
category: process_creation
product: linux
detection:
selection_parent:
ParentImage|endswith:
- '/httpd'
- '/restjavad'
- '/tmm'
- '/java'
- '/nginx'
selection_child:
Image|endswith:
- '/sh'
- '/bash'
- '/dash'
- '/python'
- '/perl'
- '/curl'
- '/wget'
- '/nc'
- '/ncat'
- '/base64'
condition: selection_parent and selection_child
falsepositives:
- Legitimate BIG-IP iApp or monitoring scripts (validate parent-child lineage before whitelisting)
- Vendor automation frameworks executing health checks
level: critical
---
title: Webshell-Style File Creation in Web-Served Directories
id: 8c2e5d41-1a6f-4e93-b7d8-2c9a4f6e1b05
status: experimental
description: Detects creation of script files in web-accessible directories, consistent with webshell deployment following exploitation of a network edge device or adjacent web server. Apply to BIG-IP syslog file events and Linux web servers in the same exposure zone.
references:
- https://www.bleepingcomputer.com/news/security/f5-releases-emergency-patches-for-critical-big-ip-rce-flaws/
- https://attack.mitre.org/techniques/T1505/003/
author: Security Arsenal
date: 2026/02/10
tags:
- attack.persistence
- attack.t1505.003
logsource:
category: file_event
product: linux
detection:
selection_path:
TargetFilename|contains:
- '/var/www/'
- '/usr/local/www/'
- '/shared/'
- '/var/tmp/'
- '/tmp/'
selection_ext:
TargetFilename|endswith:
- '.jsp'
- '.jspx'
- '.php'
- '.sh'
- '.py'
- '.pl'
condition: selection_path and selection_ext
falsepositives:
- Legitimate deployment pipelines writing to web directories (scope to exclude known CI/CD service accounts)
- Admin troubleshooting activity in /var/tmp
level: high
A note on tuning: the first rule is intentionally narrow. On a healthy BIG-IP, httpd or restjavad spawning /bin/bash or curl is anomalous, full stop. If it fires, you have a genuine investigation on your hands — that's the bar a rule should meet.
KQL — Microsoft Sentinel / Defender
BIG-IP logs reach Sentinel via Syslog or CEF (CommonSecurityLog) ingestion through a log forwarder. This query hunts for post-exploitation process execution and suspicious outbound connections reported in BIG-IP syslog streams, plus management-plane authentication anomalies:
// Hunt for suspicious process execution and outbound activity from BIG-IP appliances
// Requires BIG-IP syslog forwarding to Sentinel via Syslog or CEF connector
let Lookback = 7d;
let BIGIPHosts = dynamic(["bigip", "f5"]); // adjust to your hostname convention
union isfuzzy=true
(Syslog
| where TimeGenerated > ago(Lookback)
| where Computer has_any (BIGIPHosts) or HostIP has_any (BIGIPHosts)
| where SyslogMessage has_any ("bash", "/bin/sh", "curl ", "wget ", "chmod +x", "base64 -d", "python -c", "reverse shell")
| where SyslogMessage !has_any ("health monitor", "cron", "iControl")
| project TimeGenerated, Computer, HostIP, Facility, SeverityLevel, SyslogMessage),
(CommonSecurityLog
| where TimeGenerated > ago(Lookback)
| where DeviceVendor == "F5"
| where Activity has_any ("denied", "violation") or DeviceEventClassID has_any ("exploit", "shell")
| project TimeGenerated, DeviceProduct, SourceIP, DestinationIP, DestinationPort, Activity, Message)
| order by TimeGenerated desc
If you're also collecting management-plane auth logs, layer this in: alert on any successful login to the BIG-IP management interface or iControl REST from a source IP outside your designated admin jump hosts. On a properly architected network that list is short — three to five addresses — which makes this a near-zero-noise detection.
Velociraptor VQL
Velociraptor doesn't deploy on BIG-IP itself, but it absolutely belongs in this hunt for the Linux systems surrounding the appliance — jump hosts, automation servers, and back-end pool members that an attacker would pivot to after compromising the edge. This artifact hunts for shell processes parented to service accounts and recently created script artifacts in temp and web paths:
-- Hunt for post-exploitation indicators on Linux systems adjacent to BIG-IP
-- (jump hosts, back-end pool members, automation servers)
-- Deploy as a hunt scoped to the DMZ / server VLANs behind the appliance
SELECT Pid, Ppid, Name, Exe, CommandLine, Username, CreateTime
FROM pslist()
WHERE (
// Shells or interpreters running under service/daemon contexts
(Name =~ '^(sh|bash|dash|python|perl|nc|ncat)$'
AND Username =~ '(www-data|apache|nginx|nobody|tomcat)')
// Command lines with classic post-exploitation markers
OR CommandLine =~ '(base64 -d|/dev/tcp/|chmod \+x|curl .*\| *(sh|bash)|wget .*\| *(sh|bash))'
)
ORDER BY CreateTime DESC
Complement this with a glob() artifact pulling files created in the last 72 hours under /tmp, /var/tmp, and web root paths, filtered to script extensions — webshells and staged payloads almost always leave that artifact trail.
Remediation and Verification Script
The following Bash script is run from an admin jump host with SSH access to the BIG-IP appliance (or directly on the appliance's advanced shell). It captures current version state, checks for anomalous processes and recent file changes consistent with exploitation, and audits management-plane exposure. It does not apply patches — TMOS upgrades go through F5's standard process — but it gives you the verification evidence your change board and IR team need.
#!/bin/bash
# BIG-IP post-advisory verification and exposure audit
# Run via SSH to the appliance or from the advanced shell
# Output: timestamped report for IR/change records
REPORT="bigip_verification_$(date +%Y%m%d_%H%M%S).txt"
echo "=== BIG-IP Version and Hotfix State ===" >> "$REPORT"
tmsh show sys version >> "$REPORT" 2>&1
echo "" >> "$REPORT"
# Compare the build above against the fixed-version table in F5's advisory.
# If your branch is end-of-support, no fix is coming — escalate to upgrade planning.
echo "=== Anomalous Processes (daemons spawning shells/tools) ===" >> "$REPORT"
ps auxf | grep -Ei '(httpd|restjavad|tmm|java).*(sh|bash|python|perl|curl|wget|nc )' \
| grep -v grep >> "$REPORT" 2>&1
echo "" >> "$REPORT"
echo "=== Recently Modified Files in Temp and Web-Accessible Paths (last 72h) ===" >> "$REPORT"
find /tmp /var/tmp /var/www /usr/local/www /shared \
-type f \( -name '*.sh' -o -name '*.php' -o -name '*.jsp' -o -name '*.py' -o -name '*.pl' \) \
-mtime -3 -exec ls -la {} \; >> "$REPORT" 2>&1
echo "" >> "$REPORT"
echo "=== Established Outbound Connections from Appliance ===" >> "$REPORT"
netstat -anp 2>/dev/null | grep ESTABLISHED | grep -v -E '(127.0.0.1|::1)' >> "$REPORT" 2>&1
echo "" >> "$REPORT"
echo "=== Management Interface Port Lockdown Audit ===" >> "$REPORT"
tmsh list net self allow-service >> "$REPORT" 2>&1
echo "Review: self IPs should NOT allow tcp:443 or tcp:22 from untrusted networks." >> "$REPORT"
echo "" >> "$REPORT"
echo "=== Recent Admin Authentication Events ===" >> "$REPORT"
grep -Ei '(login|session|authenticated)' /var/log/audit* 2>/dev/null | tail -50 >> "$REPORT"
echo "Report written to $REPORT — review and attach to your change/IR ticket."
Triage guidance on the output: any process hit in section two, or any script file in section three you cannot attribute to a documented change, is a sev-1 investigation trigger, not a "monitor and see." Assume compromise until proven otherwise — the appliance handles credentials and session tokens for everything behind it.
Remediation
In priority order — and I mean this order, because exposure reduction buys you time while patching buys you safety:
-
Verify management-plane exposure now. Confirm the BIG-IP management interface (TMUI) and iControl REST are reachable only from designated admin networks or jump hosts. If you discover internet exposure, treat it as an incident, not just a misconfiguration — pull auth logs and hunt retroactively.
-
Apply F5's emergency patches per the vendor advisory referenced in the BleepingComputer report (https://www.bleepingcomputer.com/news/security/f5-releases-emergency-patches-for-critical-big-ip-rce-flaws/ — follow to F5's official advisory for the fixed-version matrix per release train). Stage through your HA pair correctly: standby unit first, verify failover, then active. Do not skip validation of virtual server availability post-failover.
-
Lock down self-IP port access. Audit
allow-serviceon all self IPs. Anything permitting management ports (443/22) from non-admin segments is an unnecessary attack surface. This is a configuration control you can apply today, independent of patch windows. -
Confirm log forwarding. Verify BIG-IP is shipping syslog (audit, LTM, ASM/AdvWAF if licensed) to your SIEM and that the KQL/CEF queries above actually return data. A patch without telemetry leaves you blind to pre-patch compromise.
-
Hunt retroactively. Run the detection content above against at least 30 days of retained logs. Emergency patches frequently follow exploitation discovery — assume the window existed before the advisory dropped. Check for unexpected config changes (
tmsh list sys last-config-changeequivalents and config sync timestamps), new local accounts, and TMSH command history anomalies. -
Handle end-of-support devices. If you're on an EoS branch, patching is not an option — only upgrading is. Escalate to leadership with the framing that this is a critical RCE on an internet-facing device with no fix path on your current version. That conversation gets budgets approved.
-
Monitor CISA KEV. If any of these vulnerabilities are added to the Known Exploited Vulnerabilities catalog, federal agencies face binding remediation deadlines, and your internal SLA should tighten accordingly — KEV inclusion is your signal that exploitation is confirmed, not theoretical.
A final word from the IR side: in previous BIG-IP compromise engagements, the organizations that fared best were the ones that could answer two questions within an hour — what is our exposure? and what do our logs show? The organizations that fared worst discovered they had neither answer and were rebuilding timelines from memory. Run the verification script. Check the logs. Patch in a controlled hurry.
Related Resources
Security Arsenal Penetration Testing Services AlertMonitor Platform Book a SOC Assessment vulnerability-management Intel Hub
Is your security operations ready?
Get a free SOC assessment or see how AlertMonitor cuts through alert noise with automated triage.