Autodesk 3ds Max CVE-2026-0537

Exploitation requires the victim to parse an attacker-supplied RGB-format image file through Autodesk 3ds Max - either by opening it directly, importing it as a texture/asset, or loading a scene that references it - meaning the attacker must deliver the malicious .rgb file to a workstation that has 3ds Max installed and convince a user (or an automated asset pipeline) to process it. Limiting factors include the Local attack vector (AV:L) confining exploitation to assets that reach the host, the practical need for user interaction to open or import the file despite the CVSS vector's UI:N marking, and the niche install base of 3ds Max (primarily 3D modeling, VFX, and architectural visualization workstations rather than general enterprise endpoints); no authentication to a network service is required because exploitation is purely client-side file parsing.

The CVSS 8.4 score reflects high confidentiality, integrity, and availability impact (C:H/I:H/A:H) with low attack complexity and no privileges or user interaction in the scoring vector, but the attack vector is explicitly Local (AV:L) - meaning the malicious RGB file must reach the workstation and be processed by 3ds Max, which in practice requires user interaction to open or import the file even though UI:N is set. Real-world risk signals are otherwise low: EPSS is 0.01% (1st percentile), no CISA KEV listing, and no public POC is identified at time of analysis, putting this firmly in the 'high CVSS, modest practical urgency' category typical of client-side file-parsing bugs. Note the apparent inconsistency between UI:N in the vector and the description's clear requirement that a user parse the file - defenders should treat realistic exploitation as user-interaction-dependent regardless of the scoring. Priority is moderate: patch on a normal cycle for design/VFX workstations, but it is not an emergency for most environments.

An attacker crafts a malicious .rgb image file with malformed header or scanline length fields and delivers it to a 3ds Max user - for example, embedded in a shared scene/asset bundle, sent via a freelance contractor handoff, or staged on a compromised asset library - and the artist opens or imports it into 3ds Max, triggering the out-of-bounds write and giving the attacker code execution as the logged-in user. No public exploit has been identified at time of analysis, so a real-world attack would require attacker-developed exploitation; the local attack vector and reliance on user file-handling make this a plausible targeted vector against VFX, architectural visualization, and game-development studios rather than a mass exploitation candidate.