The perimeter has authentication. The authentication can be spoofed. Effective today, the CVE-2026-104935 record documents a vulnerability in Ivanti Connect Secure that lets a remote, unauthenticated attacker execute arbitrary commands on the VPN appliance without providing credentials. CISA added it to the Known Exploited Vulnerabilities catalog the same day. Federal agencies have until Thursday.
The vulnerability is classified as authentication bypass by spoofing, CWE-290. It affects Ivanti Connect Secure versions before 25.1.3.0 and does not affect version 22.7R2.14 or prior. The CVSS 3.1 base score is 8.6 High, with the vector AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:L/A:L. Remote, no privileges, no user interaction. The impact profile favors confidentiality (High) over integrity and availability (both Low) — the attacker reads more than they write, which is the shape you expect when the trust boundary itself is the target rather than the service underneath it.
No workaround exists. Ivanti’s fix versions are 25.1.3.0, 25.1.3.1, and 22.8R2.5. The KEV entry requires federal agencies to apply mitigations by October 16 and flags forensics triage requirements — meaning CISA expects post-exploitation artifacts on unpatched appliances, not just a theoretical risk.
AD FS joins the pattern Forkast keeps documenting: the default was insecure, and the fix arrives later as forced remediation. Ships Without Auth covered four control planes that shipped that way. The SonicWall SMA 1000 had pre-auth SSRF. AD FS Golden SAML enforcement forced ACL hardening three months after disclosure. Ivanti’s version is the VPN perimeter itself: authentication exists, but the mechanism trusts a signal that can be forged.
The spoofing classification matters because it narrows the attack surface to identity verification at the connection layer, not a missing login screen. An attacker does not need credentials. They need the appliance to accept a forged identity claim as authentic. That is a different failure mode than SonicWall’s SSRF door or ClearPass’s missing authentication — the lock exists, but it opens for the wrong key.
CISA’s 72-hour deadline is the shortest remediation window on a VPN vulnerability in recent KEV history. The BOD 26-04 required action names two obligations: patching under the prioritized security-update guidance and forensics triage. The triage requirement signals that CISA has evidence of post-exploitation activity, not merely proof-of-concept availability. For organizations running Connect Secure as their primary remote-access gateway, that combination — active exploitation plus triage obligation — means the question is not whether to patch but whether the appliance has already been touched.
What changes between now and October 16 is mechanical. Every Connect Secure appliance running a version before 25.1.3.0 that faces the internet is a candidate for unauthenticated command execution by identity spoofing. The fix versions are published. The KEV entry is live. The forensics triage requirement is on the record.
The practical list: identify every Connect Secure deployment on the current inventory, check the running version against the unaffected range (22.7R2.14 or prior escapes the blast radius), schedule the upgrade to 25.1.3.0 or later, and treat any appliance that was internet-facing and unpatched during the disclosure window as a forensics candidate. Thursday is three days away. The authentication was not missing. It was spoofable. That distinction decides whether the fix is a login page or a firmware upgrade.