AI agent used Zammad zero-days to breach Dutch vulnerability disclosure non-profit
Attackers used an agentic AI to chain two Zammad zero-days into DIVD's network, hijacking sessions, gaining root, and exfiltrating data on September 21.
The Dutch Institute for Vulnerability Disclosure (DIVD) disclosed that a September 21 attack driven by an agentic AI exploited two Zammad zero-days to hijack sessions, execute code remotely, and escalate from the Zammad user to root, then access other services and exfiltrate data. CVE-2026-102489 is an unauthenticated RCE affecting Zammad 6.3.0-6.5.4; CVE-2026-102490 allows low-privileged authenticated attackers to reach root and affects all versions including the latest alpha; neither had a fix at publication, though CVE-2026-102489 is not exploitable in versions 7.0.0-7.1.3. DIVD reported the incident to the Dutch DPA, NCSC-NL, and police, says network segmentation limited the compromise, and advises Zammad users to upgrade to version 7 or take instances offline.
- Agentic AI attack hit DIVD on September 21 via two Zammad zero-days
- CVE-2026-102489: unauthenticated RCE in Zammad 6.3.0-6.5.4; CVE-2026-102490: root privilege escalation in all versions
- Both flaws lacked fixes at publication; users advised to upgrade to Zammad 7 or go offline
- Network segmentation and incident response limited the compromise, but DIVD assumes breach
- ATT&CK
- T1190 Exploit Public-Facing ApplicationT1068 Exploitation for Privilege EscalationT1557 Adversary-in-the-MiddleT1110.003 Password SprayingT1041 Exfiltration Over C2 Channel
- Organizations
- DIVDZammadMerlon SecurityTransluceNCSC-NL
Vulnerabilities mentioned
All| CVE | Vulnerability | CVSS | EPSS | Signals | Published |
|---|---|---|---|---|---|
CVE-2026-102489+1 related CVE | Session hijack to RCE in Zammad | 9.4 | 50% | KEVPoC ×2 | KEV |
From the original article
3 min read at helpnetsecurity.com
An agentic AI-powered attack that hit the Dutch Institute for Vulnerability Disclosure (DIVD) on September 21 exploited two zero-day vulnerabilities in Zammad, an open-source helpdesk and customer support ticketing system. “Used together, [the two flaws] allowed the attackers to hijack sessions, run code remotely and escalate privileges from the Zammad user to root, in seconds, due to the agentic part of this hack.
By Zeljka Zorz · helpnetsecurity.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.