Windows VBS Enclave CVE-2026-70075

Exploitation requires the attacker to already have authorized high-privilege local access (PR:H) to a Windows system where Virtualization-Based Security (VBS) Enclaves are in use, and to successfully win the time-of-check/time-of-use race window in the enclave's resource validation (CWE-367). Limiting factors: no remote vector (AV:L, local only), no network exploitation, high privileges already needed before the attack, non-deterministic success due to the race condition, and relevance only to hosts actually employing VBS Enclaves rather than all Windows installs. Impact is confined to integrity/security-feature bypass (C:N, A:N), not data theft or denial of service.

This is a low-to-moderate priority issue that is unlikely to warrant emergency patching but should be included in routine Windows update cycles. The CVSS 3.1 vector (AV:L/AC:L/PR:H/UI:N/S:U/C:N/I:H/A:N, base 4.4 MEDIUM) describes a local-only vulnerability requiring HIGH privileges (PR:H) that an already-authorized attacker leverages to bypass the VBS Enclave security boundary, with impact limited to integrity (I:H; no confidentiality or availability loss). All exploitation signals point the same direction and are consistent: CISA SSVC reports Exploitation=none, Automatable=no, and Technical Impact=partial, indicating no observed exploitation and no mass/automated exploitability. The CVE is NOT listed in CISA KEV and there is no public exploit identified at time of analysis. One internal inconsistency to verify: the input vector lists AC:L, but CWE-367 (TOCTOU race condition) inherently requires winning a timing window, which typically justifies AC:H - Microsoft may have rated complexity optimistically, and real-world exploitability is likely lower than AC:L implies. Because exploitation presupposes the attacker already holds high local privileges, this is a defense-in-depth / privilege-boundary weakening rather than an initial-access vector, so real-world priority is modest despite the product's ubiquity.

An attacker who already has high-privilege local code execution on a Windows host with VBS Enclaves in use races the check-to-use window of the enclave, substituting a resource after validation but before the enclave acts on it, thereby bypassing the VBS integrity protection the enclave is meant to enforce. No public exploit code is available, and the attack requires local access plus winning a non-deterministic timing race, making reliable exploitation difficult.