CVE-2026-102489, CVE-2026-102490: Zammad Helpdesk — AI-powered breach, root access (September 2026)

CVE-2026-102489, CVE-2026-102490 — Zammad Helpdesk

These two critical CVEs in Zammad Helpdesk (CVE-2026-102489, CVE-2026-102490) enable unauthenticated attackers to hijack sessions, execute arbitrary code remotely, and escalate privileges to root. Both vulnerabilities were exploited in a real-world zero-day attack on September 21, 2026, targeting the Dutch Institute for Vulnerability Disclosure (DIVD). Active exploitation is confirmed; CVSS scores are pending but expected to be 9.8+.

Attack Vector

Attackers leveraged AI-powered automation to chain the two zero-days, bypassing authentication and session controls in Zammad. The initial vector allowed session hijacking, followed by remote code execution and privilege escalation to root. The attackers exfiltrated sensitive data before being contained by network segmentation and incident response. Indicators of compromise include anomalous session tokens, unexpected root-level processes, and outbound data transfers from Zammad hosts.

Who Is at Risk

All organizations running Zammad Helpdesk are at risk, especially those with internet-exposed instances. The Dutch Institute for Vulnerability Disclosure (DIVD) is a confirmed victim. Any deployment not yet patched is vulnerable to similar AI-driven exploitation.

Patch & Mitigate

• Patch: Apply vendor-supplied patches for CVE-2026-102489 and CVE-2026-102490 immediately. Monitor Zammad advisories for release details.
• Workaround: If patching is not possible, restrict external access to Zammad servers and enforce network segmentation.
• Detect: Review logs for unauthorized session creation, privilege escalation events, and unexpected outbound connections from Zammad infrastructure.

MITRE ATT&CK

• TA0001 — Initial Access: Attackers exploited exposed Zammad endpoints to gain entry.
• TA0002 — Execution: Remote code execution was achieved post-authentication bypass.
• TA0004 — Privilege Escalation: Attackers escalated privileges to root on compromised systems.

Source: https://www.helpnetsecurity.com/2026/10/01/divd-agentic-ai-attack-breach/