Apple visionOS EUVD-2026-104696

Exploitation requires attacker-controlled code already executing on the device - i.e. a malicious or compromised app installed on an Apple Vision Pro running visionOS earlier than version 27 (PR:L reflects this low-privilege app context). No remote access and no user interaction at exploit time are needed (UI:N), and no special non-default configuration is indicated. The principal limiting factors are: the attacker must first get an app onto the headset (local vector, AV:L), impact is confined to reading sensitive user data (C:H, with no integrity or availability impact), and the issue is fixed by updating to visionOS 27.

Signals are consistent and point to a moderate, not urgent, priority. The CVSS 3.1 vector (AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N, base 5.5) describes a locally-exploitable authorization flaw requiring a low-privileged foothold - in practice, a malicious or compromised app already running on the device - with high confidentiality impact but no integrity or availability impact. This aligns with CWE-862 (Missing Authorization) and the description ('an app may be able to access sensitive user data'). The EPSS score of 0.15% (4th percentile) indicates very low probability of exploitation, and the vulnerability is NOT in CISA KEV. CISA SSVC reinforces this: Exploitation=none, Automatable=no, Technical Impact=partial. No POC is referenced. All signals agree this is a genuine but low-to-moderate issue, not a high-CVSS-but-overrated flaw nor an actively-exploited emergency. The chief caveat is exposure: exploitation presupposes attacker-controlled code already on the headset, so real-world risk is gated by the user installing a malicious app or an app being compromised. No conflicting signals were found; the main limitation is that exact affected build granularity and the specific 'sensitive user data' category are not detailed in the available data.

A user installs a malicious or trojanized app on an Apple Vision Pro running visionOS prior to version 27. Because of improper authorization state management, the app bypasses an access-control check and reads sensitive user data it should not be entitled to, all locally and without user interaction or a prompt. No public exploit code is referenced, so this requires an attacker to develop the technique independently.