Unica MaxAI Assistant CVE-2025-31992

Exploitation requires an authenticated user with at least low-level access to the HCL Unica MaxAI Assistant (PR:L), and a victim must interact with the injected content for it to render (UI:R) - these two factors are the primary limits on real-world risk. The attacker must reach an input surface within the MaxAI Assistant that processes special HTML characters client-side and renders them without neutralization; the specific feature or field acting as the injection sink is not named in the description. Impact is confined to the user's own session context (scope unchanged), limiting the ability to pivot to other users or systems.

This is a genuinely low-priority, low-risk issue despite being a client-side injection flaw. The CVSS 3.1 score of 4.6 (Medium) is driven down by meaningful limiting metrics: although the attack vector is network (AV:N) with low complexity (AC:L), exploitation requires both prior authentication (PR:L) and user interaction (UI:R), with unchanged scope (S:U) and only Low confidentiality and integrity impact and no availability impact. The EPSS score of 0.17% (6th percentile) corroborates low real-world exploitation likelihood, and there is no CISA KEV listing and no public proof-of-concept, so there is no evidence of active exploitation or weaponization. All signals are consistent and point the same direction - a legitimate but minor client-side HTML injection in a niche enterprise product, not an urgent patch-now scenario. No SSVC decision data was provided in the input, so an automatable/technical-impact determination cannot be made. The CWE-80 classification (improper neutralization of HTML tags) aligns with the XSS tag and the description; the primary unknown is the exact injection sink and whether an attacker can affect a victim other than themselves, which the scope-unchanged (S:U) metric and the phrase 'in the context of the user's session' suggest is limited.

An authenticated low-privileged user of the HCL Unica MaxAI Assistant submits input containing special HTML characters into a field that is later rendered without proper neutralization. When a user views that content, the injected markup is processed client-side within their session, potentially altering the displayed page or capturing session-context data. No public exploit code is known, and successful exploitation still requires the victim to interact with the malicious content.