Key points
- The exploit achieves code execution as root on the host system.
- Attackers hijack a live Firecracker thread from inside the guest VM.
- The method relies on modifying saved registers during KVM returns.
The team demonstrated how they manipulated register saving and resumption processes to control host execution flow. This technique allows attackers to run arbitrary code while keeping the virtual machine accessible via SSH.The team showed how to break out of a virtual machine to control the main computer.The exploit hijacks the pt_regs structure during KVM_RUN returns to redirect host execution flow.
How we got here
Firecracker is designed to isolate untrusted code within microVMs, but the researchers found a way to bypass this containment. PwnEscape Part 1 previously introduced a host memory primitive that provided access to selected host physical memory.reported The current work builds on that foundation to locate Firecracker in memory and take over a live thread.
What is new
The new report details the specific chain used to achieve CVE-2026-99999 on a Firecracker setup.reported Attackers identified the Firecracker process, found callable code within it, and located the register frame for its active virtual CPU thread. They then redirected the thread when it returned from the KVM system call to userspace.
Reading between the lines
The exploit targets the pt_regs structure, which holds saved userspace registers for the Firecracker thread while it is in the KVM system call. By modifying the instruction pointer and stack pointer in this structure, attackers dictated where execution resumed and which stack was used. A guest operation triggered the return, causing the host to execute the attacker's chosen code.
The open questions
The sources state that this exploit requires locating the Firecracker process and identifying the correct register frame using a known guest physical address. The evidence shows that attackers placed a recognizable byte pattern, or nonce, in a guest page to help resolve the translation between guest and host physical addresses.
Why it matters
- The persistence of SSH access after exploitation indicates the VM remains operational during the compromise.
- Understanding the reliance on host memory primitives helps assess the risk of other microVM escape vectors.
- The specific targeting of KVM return paths highlights a critical interface between guest and host execution contexts.
Topics: Firecracker, KVM, VM Escape, CVE-2026-99999, Security
Updated
The strongest doubt
Counterpoint from the sources
The report notes that Firecracker kept running and the VM remained accessible, suggesting the exploit does not crash the host process.
Ask the article
Local keyword search of this article; no generated answers.Did this article give you the evidence you needed?
Your choice stays on this device. No public tally.Questions readers ask
What is the CVE ID for this Firecracker escape?
The report identifies the vulnerability as CVE-2026-99999.
Does the VM stop working after the escape?
No, the VM remained accessible over SSH and Firecracker kept running.
Connected stories
Reading receipt
- The exploit achieves code execution as root on the host system.
- Attackers hijack a live Firecracker thread from inside the guest VM.
- The method relies on modifying saved registers during KVM returns.
