Ackatec
All insights

Ivanti Connect Secure Zero-Day Chain: Why You Must Patch Again and Rotate Active Directory Secrets

Ivanti's October 13 VPN fix left two exploited flaws open. Why you must upgrade again, and which Active Directory secrets to rotate and hunt for.

Author
Ackatec
Published
Reading time
6 min read
  • Ivanti Connect Secure
  • Edge devices
  • Vulnerability management
  • Active Directory
  • Incident readiness
  • Penetration testing

If you upgraded Ivanti Connect Secure after the October 13, 2026 advisory, you are not done. That release fixed one part of a three-flaw exploit chain. On October 29, Ivanti shipped fixes for the other two, CVE-2026-111357 and CVE-2026-118527, and CISA added both to its Known Exploited Vulnerabilities catalog (opens in a new tab) the next day.

The short answer: upgrade again to 25.1.1.3 or 22.8R2.6 or later, do not treat a clean Integrity Checker result as proof you are safe, and assume the attacker's real target was your Active Directory. Hunt there, and rotate the secrets your VPN appliance could reach.

What happened with the Ivanti Connect Secure zero-days?

According to Volexity's investigation (opens in a new tab), a previously unknown actor it tracks as UTA0576 began exploiting a Connect Secure VPN appliance on October 3, 2026. The appliance was fully patched at the time, on version 22.8R2.4. The actor chained three weaknesses to get unauthenticated, root level code execution:

  • An authentication bypass in a public login script that let requests reach protected scripts.
  • A file write in an upload handler that briefly left attacker files on disk before deleting them.
  • A privilege escalation in how a root level maintenance service trusts shared configuration data.

Ivanti tracks the bypass and code execution under CVE-2026-104935, fixed on October 13. The privilege escalation is CVE-2026-111357, and Rapid7 notes (opens in a new tab) that Ivanti's updated advisory names CVE-2026-118527 as an additional component of the chain. Ivanti says exploitation affected a limited number of customers.

Why was the October 13 patch not enough?

Because it closed only the entry point. Rapid7 states plainly that the October 13 fixes in 25.1.1.2 and 22.8R2.5 do not remediate CVE-2026-111357 or CVE-2026-118527. Those are the releases many teams installed in response to the first advisory, so they now need a second upgrade.

Per Rapid7's summary of Ivanti's advisory, the releases that close all three CVEs are:

  • 25.1.1 branch: 25.1.1.3 or later.
  • 22.8 branch: 22.8R2.6, or a move to 25.1.1.3 or later.
  • 25.1.3 on ISA4500, ISA6500, or ISA8500 hardware: Ivanti lists 25.1.3.0 and later as unaffected.

Ivanti recommends moving from 22.8 to 25.1.1.3 or later and using 22.8R2.6 only if migration cannot happen right away. Plan for that move regardless: CISA notes (opens in a new tab) that Connect Secure 22.8 reaches end of life on January 31, 2027. Confirm your exact version against Ivanti's current advisory before you schedule the work.

Can the Ivanti Integrity Checker detect this compromise?

No, not reliably. Volexity found that the exploit patched library code in memory only, restored the configuration it changed, and carried code built to remove matching entries from the appliance's debug and Linux audit logs. No web shell or backdoor was left behind. Each command re-ran the whole chain and returned its output in the HTTPS response. As a result, Volexity reports that Ivanti's Integrity Checker Tool did not detect any modifications, and CISA says determining compromise requires network and memory forensics.

That changes the order of operations. Rapid7 advises preserving memory and logs before rebooting or resetting an appliance where feasible. A reboot to finish an upgrade can erase the volatile evidence that would answer whether you were hit.

What did the attackers do after getting in?

They went after identity. Volexity reports that UTA0576:

  • Extracted credentials stored on the appliance, including domain credentials used for LDAP.
  • Used the appliance's Active Directory service account to search for Windows LAPS local administrator passwords, BitLocker recovery information, and password bearing attributes.
  • Ran password spraying against Active Directory accounts and tested credentials with a custom Kerberos tool.
  • Parsed packet captures on the appliance for cleartext LDAP and HTTP Basic credentials and NTLM challenge responses.
  • Collected the appliance's configuration and searched its authentication logs for valid usernames.

Volexity assessed with high confidence that this was preparation for lateral movement and a broader compromise. In other words, patching the VPN does not undo what the VPN's own account may already have read.

How do you respond if you run Ivanti Connect Secure?

1. Upgrade every appliance to a fully fixed release

Include virtual and cloud instances and any lab or disaster recovery appliances. Treat this as an emergency change, as Rapid7 recommends, not a monthly patch item.

2. Hunt on the appliance and on the network

Volexity recommends watching for large numbers of requests to the public welcome.cgi login script that carry an X-DSCGI-PATH header, alerting on any mismatch between the URL path and that header, and checking /tmp/ for post exploitation artifacts. Volexity also published related indicators of compromise that you or your incident response provider can check against appliance memory and disk images.

3. Hunt on the Active Directory side

The attacker could clean the appliance's logs, but not yours. Review domain controller and identity monitoring logs for the VPN service account reading LAPS attributes such as msLAPS-Password, msLAPS-EncryptedPassword, or legacy ms-Mcs-AdmPwd, or reading BitLocker recovery information. Look for bursts of failed authentications and unusual Kerberos service ticket requests originating from the appliance's internal address. A VPN normally looks up users and groups. It has no routine reason to read recovery secrets.

4. Rotate what the appliance could reach

If the appliance was exposed during October or any hunt finds a hit, follow the recovery steps CISA summarizes from Ivanti: back up the configuration, factory reset or rebuild, upgrade, restore, then reset passwords for all accounts, including domain, administrator, and local accounts, and revoke and reissue stored certificates, API keys, and license server credentials. Beyond that list, we recommend rotating LAPS passwords and BitLocker recovery passwords for any machine the service account could read, and, as a precaution, resetting service accounts with service principal names, since the actor's Kerberos tool included Kerberoasting functionality even though Volexity did not see it used.

5. Shrink the service account's reach

The deeper fix is least privilege. A security assessment can map what your remote access service accounts can read in Active Directory and which recovery secrets sit within reach of an internet facing device. An authorized penetration test that includes your internal network, when in scope, can show how far an attacker could move from a compromised edge device.

Questions to answer this week

  1. Which Connect Secure version is every appliance running right now, including virtual ones?
  2. Did we stop at 25.1.1.2 or 22.8R2.5 after the first advisory?
  3. Can we capture appliance memory before rebooting, and who would do it?
  4. What can our VPN service account read in Active Directory today?
  5. Would our logs show that account reading LAPS or BitLocker data?
  6. Do we have a plan and budget to leave 22.8 before January 31, 2027?

If the answers to questions 3 or 5 are unclear, an incident readiness exercise built around a compromised edge device can turn this into practiced roles and decisions. For how CISA now expects forensic triage after exploited edge flaws, see our post on BOD 26-04 and the December 7 deadline. Acting on these steps reduces risk but does not guarantee that no compromise occurred or will occur. Talk with Ackatec about where to start.

References