Nutanix Prism Element: internal management header grants unauthenticated cluster-admin JWT

Impact

Sending a crafted internal management HTTP header to Prism Element returns a Set-Cookie JSESSIONID that decodes to a JWT with maximum cluster administrative privileges, with no credentials at any point. From there an attacker executes root-level commands on every Controller VM in the cluster and reads guest VM memory and virtual disks - so every tenant VM on that cluster, including any with passthrough accelerators, is fully readable and destructible. For an operator this is a management-plane takeover, not a single-host bug: the blast radius is the whole cluster and its storage. Nutanix rates it 9.3 with exploitation reported as proof-of-concept.

Who can reach it

Anyone with network reach to the Prism Element web interface - typically the management VLAN, and the internet wherever Prism was published. Completely unauthenticated.

What to do

Apply the AOS version named in Nutanix Security Advisory 0058; the advisory PDF is the authority on fixed releases and the record does not restate them. AOS upgrades are a rolling CVM upgrade - each Controller VM restarts in turn, which is non-disruptive to running VMs but takes a maintenance window across the cluster. Until then, restrict Prism Element to the management network and block the internal management header at any reverse proxy in front of it.

References

Related entries

• KVM arm64: negative stage-1 walk level mis-sizes VNCR TLB invalidation to zeroCVE-2026-89775 · Linux kernel KVM arm64 nested virtualization (VNCR TLB invalidation size, pgshift_level_to_ttl)Critical
• KVM arm64: missing VA sign extension in range-based TLB invalidation decodingCVE-2026-89914 · Linux kernel KVM arm64 nested virtualization (decode_range_tlbi VA sign extension)Critical
• Linux kernel KVM/arm64: VM-wide VNCR mapping counter lets TLB invalidations be missed under nested virtCVE-2026-89915 · Linux kernel KVM/arm64 (nested virtualization VNCR TLB invalidation)Critical
• Linux kernel KVM/arm64: VNCR invalidation races nested page fault and installs a stale TLB entryCVE-2026-89916 · Linux kernel KVM/arm64 (VNCR fault vs. invalidation race, mmu_invalidate_seq)Critical
• Linux kernel KVM/arm64: address rollover at the end of VA space makes TLB invalidation by VA silently failCVE-2026-89918 · Linux kernel KVM/arm64 (TLBI range comparison, end of TTBR1 VA space)Critical
• Linux kernel KVM nVMX: queued TLB flushes skipped when a nested VM-Enter fails, leaving stale VPID mappingsCVE-2026-89930 · Linux kernel KVM nVMX (TLB flush on failed nested VM-Enter)Critical