In the Linux kernel, the following vulnerability has been resolved: KVM: x86: Unconditionally reset MMU context when emulating INIT Force an MMU reset when emulating an INIT even if paging was already disabled, as the MMU page role includes guest_mode (L2), i.e. an INIT while L2 is running with CR0.PG=0 will result in KVM running L1 with a stale MMU context. Alternatively, the reset could be scoped strictly to paging being enabled or L2 being active, but SMM is also encoded in the page role. Even though it _should_ be impossible for the vCPU to be INIT'd while SMM is active, there's zero value in not handling SMM. At that point, the only scenario where KVM _doesn't_ need to reset the MMU context is on an INIT with the vCPU running in Real Mode or Unpaged Protect Mode, and while INIT in those modes is somewhat common, it's far from a hot path and re-initializing the MMU is not expensive, especially when the role isn't changing (KVM elides most of the re-configuration in that case).
Casky was already ahead
This CVE exploits attack patterns that Casky's 0matched skills already investigate — long before this vulnerability was disclosed. Claude's reasoning model maps these techniques to MITRE ATT&CK, so practitioners who ran these skills have already seen the threat behaviour in their findings.
CVE-2026-99312 is a kernel-level virtualization vulnerability in KVM's x86 implementation affecting the memory management unit (MMU) context handling during CPU INIT emulation. When a virtual machine executes an INIT instruction while running in nested virtualization mode (L2 guest) with paging disabled, KVM fails to properly reset the MMU context, leaving stale L2-specific page role information active when control returns to L1. This creates a privilege escalation and guest escape scenario where an attacker with guest OS access can manipulate MMU state to access memory regions they shouldn't have permission to reach. Any organization running KVM-based virtualization infrastructure—including cloud providers, enterprise hypervisors, and local virtual machine deployments—is affected, particularly those supporting nested virtualization or running untrusted guest workloads.
While Casky.ai currently shows zero matching skills for this specific CVE, practitioners monitoring KVM hypervisor activity would detect attack patterns associated with Privilege Escalation (T1134 - Access Token Manipulation variants), Escape to Host (T1611), and Exploitation for Privilege Escalation (T1068) through extended behavioral analysis. Using Claude AI's reasoning capabilities, Casky's platform would correlate unusual MMU-related system calls, nested VM state transitions with paging mode changes, and anomalous memory access patterns that deviate from normal guest behavior. Security teams should focus detection on monitoring INIT instruction sequences in nested virtualization contexts and tracking MMU page role inconsistencies through kernel tracing, as these represent the technical indicators of exploitation attempts against this vulnerability.
Composite risk scoring from EPSS, CISA KEV, Shodan, and GreyNoise — 21 security APIs correlated into a single Casky Risk Score. Coming in Casky Pro. Join early access →
Casky has 0 skills that investigate the attack patterns behind CVE-2026-99312. Run one and get CVSS-scored findings in 3 minutes.
Run the skill that detects this →© 2026 Casky.AI, Inc. · AI Security Investigation