Recommended Posts

Ok... I got your point.. :)

If virtual machine has no network for access to your host,your host won't get affect by any virus in your guest operating system. :)

What? You're thinking of trojans, viruses don't need internet/network connections to operate.

His point was that if the guest can not access host via network, then its not possible to jump from guest to host. Which is true - guest unless setup has no access to host file system, so if no network access it should not be possible for infection to jump from guest to host.

Unless person with access to host filesystem exe something off guest file system.

His point was that if the guest can not access host via network, then its not possible to jump from guest to host. Which is true - guest unless setup has no access to host file system, so if no network access it should not be possible for infection to jump from guest to host.

Unless person with access to host filesystem exe something off guest file system.

No it's not true, viruses can exploit sandbox/VM things which is why recently as I said on the other page that hyperV, virtualbox, and all the other large-scale VM solutions with the exception of VMware had an exploit leaked that will pass from the guest to the host no problem. You've also got bugs in CPUs that can allow exploits too.

And that is a RARE specific exploit - I run vmware esxi, I have not heard of any guest to host exploits?

Could you link to these exploits from guest to host?

And along with I normally run antivirus on my windows guests. And when I use that to access questionable locations or files, I have a snapshot taken before and after done just rollback to before.

I agree no system is going to be 100% fullproof - but the odds of moving from guest to host has got to be rare. My point was that if the vm has network access or shared access to the host filesystem, then sure it would be easy for something to jump to another box on the network if a worm and other boxes are open to it, and or if guest has write access to host filesystem it could infect/encrypt/delete/etc files on the host filesystem.

Ok - 2nd one is dos attack, not a guest-host escape

"This is a denial of service vulnerability. An attacker who exploited this vulnerability could cause the affected Hyper-V server to stop responding, requiring a restart. Note that the denial of service vulnerability would not allow an attacker to execute code or to elevate their user rights on the Hyper-V server, but it could cause the affected system to stop accepting requests."

3rd one again is talking about dos, not escape.

Again not saying they don't exist - there was Cloudburst back in 2009. But do you know of any active guest-host escapes in the wild? Other than whitepapers discussing the possibility?

Are you aware of any malware/extortionware/virus that uses such complicated exploits?

Again I will agree that nothing is 100% secure, and yes there always going to be security issues that need to be addressed. But it comes down to is the exploit in the wild? Is it something that you surfing the underbelly of the net, or driveby or even running some keygen or hacked game or application is going to be infecting your machine and jumping to your host?

I would be more concerned with running bad code (worm) that then could then seek out your other vms or physical or even host machine via network than a guest-host escape.

Generally speaking, not talking a targeted attack against your infrastructure - some paid hacker getting into a company through a vm that is exposed to the public net, etc. Just generally speaking its unlikely to have to be worried about a guest-host escape no matter what your running be it virtualbox, vmware workstation (type2) or something like esxi/xen/kvm (type1).

Should you be aware that such things can exist - sure! If your tinfoil hat is a bit tight, maybe you don't run the vm tools on your vm - this is generally going to be the attack vector currently. Must of the stuff I have seen has been against intel, so run amd cpus on your host ;)

But common sense would tell you not to allow your VMs to have rw to your host filesystem via vm sharing tools or setup, etc. And if your playing with bad code - you might want to isolate their network connectivity to the rest of your vms/host/local network.

I don't know of any publically available viruses that utilise it no, but the reason for that is because the people that know about these security flaws will be governments and incredibly sophisticated hackers, they won't be telling OEMs about them nor releasing POCs about them either but will use them discreetly in the shadows without anyone knowing.

"about these security flaws will be governments and incredibly sophisticated hackers"

So you think a virus scanner you got from mcafee or nod or whatever is going to protect against a government backed attack?

So as I was saying, generally speaking guest-host escapes is not something the general user base - say running some vms on my home lab box for example needs to be worried about protecting against. Other than making sure your VM software be it type1 or 2 is updated, and host os are patched and securely configured.

And in this context you could be pretty sure that this form of exploit is not going to be used when you download that keygen or hacked exe of some game or application or surf the dark underbelly of the net. What you should be concerned with is more that you can roll the vm back after you play with something like that. Or that the box does not become infected with the lastest worm that could infect your other vms/local network.

I think we are the same page yes?

Hello,

There was a discussion not too long ago of the Windows version of OSX/Crisis infecting virtual machines, and I though the Koobface malware contained an exploit for VMware that allowed it to escape from the guest OS to the host OS, but I cannot seem to find an exact reference to it. I think this would have been around 2009-2010, though.

Regards,

Aryeh Goretsky

The recent one was fine if you use vmware, it was not vulnerable.

Here's some links to things about the bugs;

http://www.aidanfinn.com/?p=12837

http://technet.micro...lletin/ms11-047

http://blog.coresecu...ploit-ms10-102/

Hello,

I am unsure of why McAfee or ESET of any of the other anti-malware programs would not protect you. They detect various pieces of malware such as ACAD/Medre, OSX/Lamadai, Win32/Duqu, Win32/Flamer, Win32/Georbot, Win32/R2D2 and Win32/Stuxnet which are generally acknowledged to be created (or tacitly approved by) various nation-states around the globe.

While it seems ludicrous that an anti-malware company might be fettered by the intelligence apparatus of the country in which it is headquartered, even if you were not to rule that out of hand, there are anti-malware companies located in countries around the world that are mutually antagonistic towards each other, and I could see a Chinese or a Russian anti-malware company treating malware created by an American or European government as a PR goldmine. The truth of the matter, though, is that governments are going to be pretty unlikely to inform anti-malware companies of malware they have developed or are using, since that violates the whole point of using malware in the first place: Plausible deniability.

Regards,

Aryeh Goretsky

"about these security flaws will be governments and incredibly sophisticated hackers"

So you think a virus scanner you got from mcafee or nod or whatever is going to protect against a government backed attack?

So as I was saying, generally speaking guest-host escapes is not something the general user base - say running some vms on my home lab box for example needs to be worried about protecting against. Other than making sure your VM software be it type1 or 2 is updated, and host os are patched and securely configured.

And in this context you could be pretty sure that this form of exploit is not going to be used when you download that keygen or hacked exe of some game or application or surf the dark underbelly of the net. What you should be concerned with is more that you can roll the vm back after you play with something like that. Or that the box does not become infected with the lastest worm that could infect your other vms/local network.

I think we are the same page yes?

This topic is now closed to further replies.
  • Posts

    • Microsoft finally admits its default Windows 11 25H2, 24H2 action broke key legacy component by Sayan Sen Microsoft last week released Windows 11 KB5094126 and KB5093998 as the latest Patch Tuesday updates. Following that the company also published the accompanying dynamic updates under KB5094149, KB5095971, and KB5094156. So far the company has acknowledged two known issues that have popped up after the release which include bugged-out Office apps as well as the Recycle Bin; though there could be more at play too. Speaking of bugs and issues, Microsoft seems to have finally acknowledged a problem that probably has been around for close to a year. That's because back in July of 2025 the company made a default change to the latest Windows 11 versions, wherein it switched to JScript9Legacy on Windows 11 24H2 and later releases. Hence following the release of version 25H2 in October 2025, JScript9Legacy also remained default-enabled. As a result there has been a compatibility issue ever since then. For those wondering, by switching to JScript9Legacy Microsoft intended to improve the security of modern Windows PCs by reducing vulnerabilities tied to legacy scripting like cross-site scripting (XSS), among others. XSS exploits can allow cyber-attackers to attach malicious code onto legitimate websites and use them to execute the code when a potential victim loads such a website. Hence the new JScript9Legacy engine enforced stricter execution policies and improved object handling, which should help mitigate such attacks. Microsoft today has published a new support article detailing the problem. Neowin spotted it while browsing. The company says that JScript global definitions and execution context may fail to persist across scripts, potentially breaking older dependent apps and web-based components that relied on this legacy behavior. In the article Microsoft has confirmed that the issue stems from its move away from the older jscript9.dll engine in favor of jscript9legacy.dll. As mentioned above, while the newer engine was designed to address vulnerabilities and strengthen security it also changes how JScript handles execution context. As a result functions and definitions loaded by one script could no longer remain available to subsequent scripts once execution ended. The company notes that some applications worked correctly on earlier Windows versions because the older JScript engine automatically retained global definitions and execution state between scripts. Under the newer model though that behavior is disabled by default causing certain legacy workloads and polyfill-dependent scripts to fail. Microsoft says it addressed the problem via the KB5077241 update though the fix had not been enabled automatically in the following updates. As such admins must explicitly turn on persistent JScript execution context using a Registry setting that the tech giant shared today. The configuration can be applied to individual processes or system-wide through the FEATURE_ENABLE_PERSISTENCE registry key. The steps have been outlined below: Run the following command to create the feature control registry key: reg add "HKLM\Software\Policies\Microsoft\Internet Explorer\Main\FeatureControl\FEATURE_ENABLE_PERSISTENCE" Under this key, create a new DWORD (32-bit) value. Configure the value as follows: To enable persistence for specific processes only: Set the value to 1 for each target process name. To enable persistence for all processes: Add * as the key name and set its value to 1. You can find the official support article here on Microsoft's website.
    • The possibility that milk gathers back into a glass implies that gravity can be 'reversed'.
    • VidCoder 12.20 by Razvan Serea  VidCoder is a DVD/Blu-ray ripping and video transcoding application for Windows. It uses HandBrake as its encoding engine. Calling directly into the HandBrake library gives it a more rich UI than the official HandBrake Windows GUI. VidCoder can rip DVDs but does not defeat the CSS encryption found in most commercial DVDs. You’ll need the NET 8 Desktop Runtime. If you don’t have it, VidCoder will prompt you to download and install it. The Portable version is self-contained and does not require any .NET Runtime to be installed. You do not need to install HandBrake for VidCoder to work. Feature list: Multi-threaded MP4, MKV containers Completely integrated encoding pipeline: everything is in one process and no huge intermediate temporary files H.264, H.265, MPEG-4, MPEG-2, VP8, Theora video Hardware-accelerated encoding with AMD VCE, Nvidia NVENC and Intel QuickSync AAC, MP3, Vorbis, AC3, FLAC audio encoding and AAC/AC3/MP3/DTS/DTS-HD passthrough Target bitrate, size or quality for video 2-pass encoding Decomb, detelecine, deinterlace, rotate, reflect, chroma smooth, colorspace filters Powerful batch encoding with simultaneous encodes Customizable Pickers to automatically pick audio and subtitle tracks, destination, titles and more Instant source previews Creates small encoded preview clips Pause, resume encoding VidCoder 12.20 changes: Updated HandBrake core to 1.11.2. Download: VidCoder 12.20 | 47.0 MB (Open Source) Download: Portable VidCoder 12.19 | 89.3 MB Link: VidCoder Home Page | Github | Screenshot Get alerted to all of our Software updates on Twitter at @NeowinSoftware
    • Too soon, I'm still not over this death!
  • Recent Achievements

    • Week One Done
      Jordan Smith earned a badge
      Week One Done
    • Reacting Well
      BizSAR earned a badge
      Reacting Well
    • First Post
      AndreaB earned a badge
      First Post
    • Week One Done
      Huge Trailer earned a badge
      Week One Done
    • Week One Done
      Classifyskilleducation earned a badge
      Week One Done
  • Popular Contributors

    1. 1
      +primortal
      593
    2. 2
      +Edouard
      185
    3. 3
      PsYcHoKiLLa
      77
    4. 4
      Michael Scrip
      73
    5. 5
      Steven P.
      66
  • Tell a friend

    Love Neowin? Tell a friend!