Skip to main content
All incident reports
Incident report

F5 patches BIG-IP APM zero-day exploited for unauthenticated remote code execution

F5 has released hotfixes for a critical BIG-IP Access Policy Manager flaw, CVE-2026-94127, which it says has been exploited to run code without authentication on systems acting as OAuth authorisation servers. CISA added the flaw to its Known Exploited Vulnerabilities catalogue the same day.

· First published

The short version

  • The flaw. CVE-2026-94127 is a heap-based buffer overflow in F5 BIG-IP APM that lets an unauthenticated attacker run code on the appliance. It affects only virtual servers configured with both an APM access policy and an OAuth authorisation server profile.
  • Exploitation. F5 says it learned the flaw has been exploited and that it found the defect internally. It has published three indicators of compromise and an iRule mitigation for customers who cannot install the hotfixes immediately.
  • Federal deadline. CISA added CVE-2026-94127 to its Known Exploited Vulnerabilities catalogue on 22 September and gave federal civilian agencies until 25 September to apply F5's mitigations.
  • Scale is unclear. Shadowserver tracks more than 14,700 IP addresses with BIG-IP APM fingerprints, but there is no information on how many are patched or honeypots, and no source says how many systems were attacked.

Timeline

DateEvent
22 SepF5 publishes an advisory for CVE-2026-94127 and releases engineering hotfixes; CISA adds the flaw to its Known Exploited Vulnerabilities catalogue.
23 SepF5 updates its CVE record at 00:45 UTC to state that the flaw is present only when APM acts as the OAuth authorisation server.
25 SepDeadline CISA set for federal civilian agencies to apply fixes or mitigations.

What happened

F5 disclosed CVE-2026-94127 in an advisory published on 22 September and released engineering hotfixes for three release branches. The company says it discovered the flaw internally and that it has been exploited. F5 describes the bug as a heap-based buffer overflow rated 9.8 on CVSS v3.1 and 9.3 on v4.0, allowing unauthenticated attackers to achieve remote code execution.

The vulnerable setup is specific: a virtual server that carries both a BIG-IP APM access policy and an OAuth authorisation server profile. Malicious traffic sent to that virtual server is enough to trigger the flaw. F5 says the problem sits in the data plane, so restricting access to the BIG-IP management interface does not protect a system, and systems running in Appliance mode are also vulnerable.

CISA added the flaw to its Known Exploited Vulnerabilities catalogue the same day and set federal civilian agencies a deadline of 25 September. Internet threat monitoring non-profit Shadowserver currently tracks over 14,700 IP addresses with BIG-IP APM fingerprints, but there is no information on how many of those have been patched or are honeypots.

Who is affected

F5 lists BIG-IP APM versions 21.1.0, 17.5.0 to 17.5.1 and 17.1.0 to 17.1.3 as vulnerable when APM acts as an OAuth authorisation server. Hotfixes exist for each branch; the fixed build for 17.5 is Hotfix-BIGIP-17.5.1.9.0.160.12-ENG, and F5's advisory lists the rest. Deployments that use APM only as an OAuth client or resource server are not affected, and F5 says no other products are vulnerable.

Two caveats matter for anyone checking their estate. F5 did not evaluate versions that have reached End of Technical Support, so their status is unknown rather than safe. And the fixes for an earlier APM flaw, CVE-2025-53521, added to the same CISA catalogue in March, fall inside the affected ranges above, so a system updated to those builds still needs the new hotfix.

The description of the trigger condition differs between publications. F5 amended its CVE record at 00:45 UTC on 23 September to limit the flaw to the authorisation server role; CISA's catalogue entry and an advisory from CERT-EU, the cybersecurity service for EU institutions, were published before that change and describe the condition more broadly, as an access policy and an OAuth profile on a virtual server.

Detection and mitigation

F5's mitigation for organisations that cannot patch immediately is an iRule, a small traffic-handling script obtained by opening a ticket with F5 support, applied to the affected virtual server. CISA told agencies to apply the iRule first to allow forensic triage and then install the vendor hotfix. CERT-EU advises preserving forensic evidence before patching and starting incident response if signs of compromise are found.

The indicator to act on is a combination, according to F5: repeated OAuth authentication failures, then suspicious commands, then a TMM SIGABRT shortly afterwards. In practice that means looking in /var/log/apm for failed UserInfo requests reporting "The access token is invalid", especially ten or more from one IP address in a short window, checking for an unexplained rise in total_failed with tmctl global_oauth_stat, and reviewing /var/log/audit around those times. TMM core files alone are not proof.

F5's record, CISA and CERT-EU do not say whether installing the hotfix removes access an attacker has already obtained. F5 vulnerabilities have a history of being targeted: CISA has flagged eight actively exploited F5 vulnerabilities since November 2021, four of which were also abused in ransomware attacks. Separately, F5 disclosed in October 2025 that state-sponsored hackers had breached its systems in August 2025 and stolen BIG-IP source code; no source connects that intrusion to this flaw.

What is not known yet

  • Who is behind the exploitation and which organisations were targeted.
  • How many exposed systems have been attacked or compromised.
  • Whether installing the hotfix removes access an attacker already gained.
  • How many of the 14,700-plus fingerprinted IP addresses run the vulnerable authorisation server configuration.
  • Whether BIG-IP APM versions past End of Technical Support are affected.

What to do

If you run BIG-IP APM

  1. Check whether any virtual server combines an APM access policy with an OAuth authorisation server profile; if none does, F5 says the system is not affected.
  2. Install the engineering hotfix for your release branch from F5's advisory, and do not assume earlier fixes for CVE-2025-53521 cover this flaw.
  3. If you cannot install the hotfix now, open a ticket with F5 support and apply the iRule mitigation to the affected virtual server.
  4. Do not treat restrictions on the BIG-IP management interface as a mitigation; F5 says the flaw is reached through the virtual server in the data plane.

If you suspect compromise

  1. Preserve forensic evidence before applying the hotfix, as CERT-EU advises.
  2. Search /var/log/apm for repeated failed UserInfo requests reporting an invalid access token, particularly ten or more from a single IP address in a short period.
  3. Check tmctl global_oauth_stat for an unexplained rise in total_failed and review /var/log/audit for suspicious commands at those times.
  4. Treat repeated OAuth failures followed by suspicious commands and a TMM SIGABRT as a trigger for incident response.

For everyone else

  1. Inventory internet-facing access and identity proxies and confirm which role each one performs, since only systems acting as an OAuth authorisation server are exposed here.
  2. Track vendor advisories for exploited flaws in the appliances you expose, and patch or mitigate them before reviewing logs.

Update log

  • First published.

Sources

New incident reports by email

One email a week with the reports published since the last one.

Want us to look at your domain?

Give us one apex domain and we'll send back a written snapshot of what is already visible from outside. No call required.