Windows Win32k CVE-2026-70285

Exploitation requires the attacker to already have local, authenticated access to the target Windows system with at least low-level privileges (PR:L) - for example an interactive or service account able to run arbitrary user-mode code - in order to invoke the vulnerable Win32k kernel subsystem path and trigger the heap overflow. No user interaction by a victim is needed (UI:N) and no special non-default feature must be enabled; the limiting factors are the need for prior local code execution (no remote/network vector, AV:L) and existing valid credentials or a foothold. Affected builds span Windows 10 (1607, 1809, 21H2, 22H2), Windows 11 (23H2, 24H2, 25H2, 26H1, 26H2) and Windows Server 2016/2019/2022/2025 including Server Core installations below the listed patched build numbers; systems already updated to the fixed builds are not exploitable.

This is a credible but not emergency-tier privilege-escalation issue. The CVSS 3.1 vector (AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H, base 7.8 HIGH) describes a local attack requiring low privileges already on the host, with full confidentiality, integrity and availability impact - i.e. a standard local-user-to-SYSTEM/kernel escalation. CISA's SSVC decision data reinforces a measured priority: Exploitation=none (no observed or public exploitation), Automatable=no (cannot be reliably mass-automated, consistent with a local-only vector that needs prior foothold), but Technical Impact=total (full kernel compromise if exploited). No EPSS score was supplied in the input, and the CVE is NOT on CISA KEV, so there is no signal of active exploitation. No POC was provided. The signals are internally consistent - local vector, prior authentication required, no automatability, no exploitation evidence - meaning this is a real patch-soon issue (kernel EoP on a ubiquitous platform) rather than a drop-everything internet-facing threat. The main caveat: as a Win32k kernel bug on Windows, this class is historically a favorite for exploit development and chaining after initial access, so defenders should not dismiss it despite the 'none' exploitation status. EPSS is missing and cannot be used to refine the probability.

An attacker who has already gained low-privileged code execution on a Windows host (via phishing, a malicious installer, or a compromised standard user account) runs a crafted local program that triggers the heap-based buffer overflow in the Win32k kernel subsystem, corrupting kernel heap structures to execute code at kernel/SYSTEM level and take full control of the machine. No public exploit was identified at time of analysis, and the attack requires local access with existing low privileges, so it is a post-compromise escalation step rather than an initial entry point.