In the Linux kernel, the following vulnerability has been resolved: KVM: arm64: Bound used_lrs when flushing the pKVM hyp vCPU flush_hyp_vcpu() copies the host vGIC state into the hyp's private vCPU on every run. The vGIC list register save and restore use used_lrs as their loop bound and expect it to stay within the number of implemented list registers. While this is generally the case, flush_hyp_vcpu() copies vgic_v3 verbatim and does not enforce this, so a value provided by the host is used at EL2 to index vgic_lr[] and access ICH_LR<n>_EL2 (host -> EL2). Fix by clamping used_lrs to the number of implemented list registers after the copy, as the trusted path already does in vgic_flush_lr_state(). The number of implemented list registers is constant after init, so it is replicated once from kvm_vgic_global_state.nr_lr into hyp_gicv3_nr_lr rather than read on every entry.
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-2026-64287 is a memory safety vulnerability in the Linux kernel's KVM hypervisor implementation for ARM64 architecture. The vulnerability exists in the flush_hyp_vcpu() function, which copies host vGIC (Virtual Generic Interrupt Controller) state into the hypervisor's private vCPU without validating critical boundary conditions. Specifically, the used_lrs (used list registers) field is copied verbatim from host-controlled data without enforcement against the actual number of implemented list registers, creating an out-of-bounds condition. This affects Linux systems running KVM with ARM64 support, particularly those hosting virtual machines where a malicious or compromised host could exploit this to corrupt hypervisor memory, potentially breaking isolation between VMs or between VM and host.
While CVE-2026-64287 lacks explicit MITRE ATT&CK technique mapping, Casky's 754 security skills powered by Claude AI with extended reasoning would identify attack patterns consistent with Memory Corruption (CWE-119) and Privilege Escalation (MITRE T1134) techniques. A practitioner using Casky would see detections focused on: (1) unsafe kernel memory operations without bounds checking, (2) host-to-hypervisor data flows that bypass validation, and (3) register allocation logic that trusts untrusted input. The findings would flag the lack of input validation between trust boundaries and highlight how an attacker could leverage this for hypervisor escape or VM isolation bypass, enabling lateral movement between virtualized workloads or host compromise.
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-2026-64287. Run one and get CVSS-scored findings in 3 minutes.
Run the skill that detects this →© 2026 Casky.AI, Inc. · AI Security Investigation