CVE-2026-99312: KVM Guest-to-Host Escape

(terra.security)

122 points | by Kozmik1 2 hours ago

17 comments

  • praseodym 1 hour ago
    The blog post is very difficult to read, I suspect that unfortunately it’s fully written by an LLM.
    • xlayn 1 hour ago
      doesn't change the value of what's writen, if true and actionable that's all that should be needed
      • manbash 25 minutes ago
        It has "potential", not "value".

        Also in this case, a "missed potential" if it eventually fails to reach anyone.

      • sylos 37 minutes ago
        Value is more than just iron something is made from. How it's shaped and presented is just as important
        • tptacek 21 minutes ago
          No, it is not "just as important" as the underlying information.
          • tsunagatta 7 minutes ago
            It actually might be far more important. You are a lot more important than the molecules that you are made of
      • drum55 55 minutes ago
        If it was valuable a human would have been able to communicate it in a few lines of text. Using a LLM to blast slop devalues everything and communicates little. A drunken twitter rant has more ability to convey information than sloppy blog posts with 3 pieces of information per sentence and a load bearing attitude.
        • Ylpertnodi 19 minutes ago
          Get hip to it, daddy-o.
        • cluckindan 48 minutes ago
          Just ask the clanker to read it buddy, you’re already behind /s
      • catlifeonmars 28 minutes ago
        It kind of does, the OP could have just saved themselves the tokens and posted a 130 word summary on x/twitter that would have been just as useful.
      • InsidiousWombat 16 minutes ago
        well it's 404 now
    • Betelbuddy 1 hour ago
      The LLMs find the security issues, the LLMs will write the reports and the LLMs will fix and patch them.

      I guess its time for all of us to step aside, leave space for somebody who knowns what they are doing, and go fishing...

      • praseodym 1 hour ago
        And this way only LLMs can learn from these vulnerabilities, making it even more likely that the next one will be found by an LLM.
        • catlifeonmars 26 minutes ago
          LLMs currently don’t learn, at least in the way most people mean, but I assume you mean the trainers will incorporate the knowledge in future releases.
    • biotechbio 1 hour ago
      Glad to see this is the top comment. This is written by Claude, I know because I talk to Claude for 8+ hours a day, and it says stuff like this. I stopped reading the article because of it.
  • jacobgold 1 hour ago
    Our infrastructure will become so much more secure but we're in for a bumpy ride through 2027.
    • dvh 1 hour ago
      Not necessarily. Only if you assume that no new insecure code is added. If ai found 1 security bugs in old code but ease of code generation with ai adds 2 new security issues (in unrelated code), then we are getting less secure.
      • tyingq 1 hour ago
        Is it a given that it's as good at outputting secure code as it finding insecure code? That leaves a fair amount of room for people not baking in deep review steps. Also mentioned elsewhere, but if velocity is way up...that ramps up bugs anyway.
        • MajesticHobo2 1 hour ago
          No, it's not a given. AI always finds security issues in the things I just had it write.
          • hparadiz 46 minutes ago
            This is true for human written code too. This is why in the future all written code MUST be run through an internal security sweep before being given to your harness for writing. That's just basic due diligence at this point.
            • vbezhenar 8 minutes ago
              As long as model providers deny access to security-capable models, it can't be basic diligence.
              • hparadiz 4 minutes ago
                This is something agents already do as part of their normal work. Providing code that you know is vulnerable is in fact a dereliction of duty and negligence. Don't provide a POC. Simply that as it produces code a background agent just checks it without showing the work and points out potential problems.
            • catlifeonmars 25 minutes ago
              Yes but LLMs famously produce vastly more code than humans. My hot take: LLMs will produce more insecure code than they can catch, and will also find ways to smuggle insecure code to avoid extra work.
      • SOLAR_FIELDS 1 hour ago
        Do we have any evidence around the rate that AI emits security bugs vs humans previously? You would think that perhaps AI would be better about this, but perhaps not.
        • dvh 1 hour ago
          I'm making no claims. I'm just saying that contributing before ai was difficult so people rarely did it. Now that contributing is easy, more code will be added. More code = more bugs. Always.
          • voakbasda 1 hour ago
            Yet the opposite seems to be true. Many projects now shun new contributors due to AI.
            • punchmesan 38 minutes ago
              Eh I'm not sure that's an indicator that the opposite is true. Github's vastly increased commit activity is the best indicator we have of this. https://www.tomshardware.com/software/github-report-highligh...

              Sure, some projects are trying to be responsible or are trying to keep waves of slop PR's from drowning them, but that is by no means an indication of a general distaste for AI code. It's been largely embraced.

            • hparadiz 45 minutes ago
              I've run Daybreak against my human written pre-AI code and found so much stuff that I don't trust human or AI written code. It all MUST be audited by a security sweep AI.
        • the8472 1 hour ago
          Even if Fable and Astra are superhuman at identifying and fixing bugs that does nothing when projects get bugslopped faster and cheaper by some low-end model.
      • tptacek 1 hour ago
        I hate these meta conversations on crunchy technical topics and feel like we should be digging into the KVM security model more than this general security prognostication stuff. Having said that, I will hypocritically further this debate by pointing out that the determinant of software security moving forward is not so whether AI is drastically better than humans at avoiding common vulnerability patterns, (or instead just a little better). Rather, it's that AI will make it much easier to move subsystems or whole projects to memory-safe programming languages and towards formal verification.
      • j45 1 hour ago
        Existing code will have security holes that haven't been found yet.
    • InsidiousWombat 15 minutes ago
      ...if the pace of both fixing and finding bugs got accelerated, all we're in for is machines being on average more vulnerable (because patching takes time and effort and more bugs increases that)
    • TZubiri 22 minutes ago
      Have you ever been told that you are problematically optimistic?
  • gerdesj 56 minutes ago
    404
  • fwsgonzo 1 hour ago
    Don't disable nested, it's still better security than everything else, and having it on means access to real hardware virtualization. Second, the article is AI-written. It really should be rewritten by a human.
    • xmodem 1 hour ago
      Outside of development and CI, I have yet to hear a compelling use case for nested virtualization.

      Agree that the article is downright painful to read though.

      • JoshTriplett 1 hour ago
        > I have yet to hear a compelling use case for nested virtualization.

        AWS instances are slow to launch; they take whole seconds. Nested virtualization allows running something like firecracker/crosvm/rust-vmm atop a normal AWS instance, without needing a bare-metal one.

        • tptacek 1 hour ago
          If you're segregating internal workloads, by all means use nested virtualization on EC2 or DO or something. Your threat model is different and you probably have a lot more room to maneuver when a KVM escape is announced, because you have some level of trust in all your workloads.

          If you're running multitenant for mutually-untrusting customers, that excuse goes out the window.

          • JoshTriplett 1 hour ago
            One distinction I'd make as well is "are you running more than one customer's workload on the same system" (yikes, don't), vs "are you using nested virt as additional protection rather than running the customer workload directly on an instance, but one instance only gets jobs from one customer".
            • necovek 23 minutes ago
              Does this hold for the likes of Amazon?

              Do your lambda functions always get a per-customer dedicated instance or is firecracker distributing them so that multiple customers can have VMs on the same actual box?

              Sounds like it would be pretty expensive for Amazon to dedicate an instance to every customer, especially for those with small workloads.

              If that's the case, why would it not hold for Amazon but would hold for everyone else?

              • swiftcoder 18 minutes ago
                > Do your lambda functions always get a per-customer dedicated instance

                At some point in the past they did, for pretty much exactly this reason. I'd assume they don't anymore (I last worked there a decade ago).

            • tptacek 1 hour ago
              Another good point, but this one is maybe a little less dispositive than it looks like: it's very difficult to engineer an orchestration system that reliably segregates worker physicals from each other, so I'd rate the likelihood of a game-over situation after a guest->host escape as very high regardless of how yuo schedule workloads. It's sort of the same situation as how SSRF's are usually game-overs (you're pivoting from a hardened attack surface to a much softer one) but even harder to engineer again.
              • JoshTriplett 57 minutes ago
                That's fair. You'd want a system designed around the notion that every customer instance must be isolated from each other without the nested virtualization.

                (Aside: my current guess at the appropriate security posture towards Nitro Sandboxes, in the current era of vulnerabilities, is "decent, but probably has security holes too, and is less widely scrutinized, don't count on it either". Thoughts?)

                Any recommendations for good guides on that, besides "design very careful IAM policies and security groups, harden any common services they have access to, and ideally run them under separate AWS accounts"?

    • TZubiri 21 minutes ago
      ? If you don't use it, disable it. Simple as that
    • rootnod3 1 hour ago
      > It really should be rewritten by a human.

      That does not seem to be what HN has been thinking for months.

  • cloudie78 6 minutes ago
    Nested virtualisation.

    Just don’t.

  • arde 51 minutes ago
    That's a 404 now.
  • gbrindisi 1 hour ago
    Thinking about agents sandboxes... it truly seems like the time is ripe for unikernels
    • wmf 29 minutes ago
      Unikernels that run on KVM aren't going to protect against KVM escapes.
      • rvz 4 minutes ago
        Maybe the problem...is Linux itself.
  • adverbly 59 minutes ago
    Well this sucks. So this needs a kernel fix then? Looks like I can disable nested passthrough manually, but am I stuck waiting for downsteam fixes at this point then?

    Honestly, this gives me pause on the whole "VM is sufficient for any agentic sandbox environment" theory.

  • tptacek 1 hour ago
    Another nested KVM bug. We debated enabling nested KVM for many years, and held the line every time, because the only post-2020 KVM vulnerability we knew of required nested VMs. That decision aged well: we've gotten at least 2 more nested KVM escapes in the last year.

    Don't allow nested KVM.

    • rwmj 1 hour ago
      Certain Windows features need it unfortunately. Virtualization-Based Security (VBS), Credential Guard, Hyper-V (obviously), and I'm going to say WSL2(?)
      • tptacek 1 hour ago
        Right, sorry, that's a fair point (and Josh Triplett made another fair point earlier). I'm sort of implicitly locked in to the "hosting company" use case here: services running mutually-untrusting isolated workloads for customers.
  • LelouBil 2 minutes ago
    [delayed]
  • Ungh 50 minutes ago
  • ArcHound 1 hour ago
    Surely, we'll run out of critical CVEs in open-source someday. Seems it's not today.

    In the meantime the patching will continue until security improves!

    • j45 1 hour ago
      Doubt it, but there will be a big spike for a while and then slow down.
  • bigbuppo 30 minutes ago
    Looking forward to Ubuntu finally rolling out an updated kernel for this in the year 2057.
  • nubinetwork 1 hour ago
    I guess its time to shut off the internet.
  • calvinmorrison 1 hour ago
    what is the mitigation.
    • stilldavid 1 hour ago
      Look in the article under the headline "Mitigation".
      • pixl97 46 minutes ago
        Page has gone 404.
    • j45 1 hour ago
      The cloud is somebody else's computer.

      Mitigations are layered, first not using a shared VPS, and instead running your own bare metal server with something like proxmox to run the containerization.

      Those containers could be susceptible to issues between them too, but you at least control the environment and can put more checks and balances in place.

      Shared cloud and vps is going to become bumpy and more insecure the next little while. There's no way around it.

  • FLeXMurphy 25 minutes ago
    Page is now 404'd.
  • k0morebi 55 minutes ago
    404 now
    • pixl97 47 minutes ago
      Confirmed, wonder what happened here.