Sef AI Chatbot EUVD-2026-91324

Exploitation requires an authenticated low-privileged account on the Sef AI Chatbot Platform (PR:L) and succeeds against versions before 2.1; the flaw lets such a user reach functionality not properly constrained by ACLs, yielding confidentiality impact only (no integrity or availability effect). Limiting factors: valid credentials are mandatory (no unauthenticated access), and the CVSS high attack complexity (AC:H) implies the attacker must satisfy additional non-trivial conditions - such as specific request state, timing, or knowledge of protected object identifiers - rather than a single trivial request. The exact ACL-gated feature is not specified in available data; upgrading to 2.1 or later remediates it.

This is a moderate-severity (CVSS 5.3) Missing Authorization flaw (CWE-862) in HAVELSAN's Sef AI Chatbot Platform. The CVSS vector (AV:N/AC:H/PR:L/UI:N/S:U/C:H/I:N/A:N) indicates a network-reachable issue that nonetheless requires an authenticated low-privileged account (PR:L) and high attack complexity (AC:H), with impact limited to confidentiality (C:H) - an authenticated user can reach data/functions that should be gated by ACLs. No EPSS score, KEV listing, or POC was provided in the input, so there is no signal of active exploitation or weaponization; this reads as a real-but-not-urgent authorization defect rather than a mass-exploitation threat. The combination of required authentication, high complexity, and a niche vendor product keeps real-world risk well below what the 'network vector' alone might suggest. Note a minor signal conflict: the 'Authentication Bypass' tag is imprecise - CWE-862 and the description ('functionality not properly constrained by ACLs') describe an authorization/access-control bypass by an already-authenticated user, not an authentication bypass.

An attacker who already holds a valid low-privilege account on the Sef AI Chatbot Platform (e.g., a standard chatbot user) crafts requests to API endpoints or functions whose access is not properly restricted by ACLs, retrieving data or invoking capabilities intended only for higher-privileged roles. The high attack complexity suggests non-trivial conditions (specific request sequencing, state, or knowledge of internal object references) must be met. No public exploit code was identified at time of analysis.