I work on a corporate laptop that has an infamous root CA certicate installed, which allows the company to intercept all my browser traffic and perform a MITM attack.
Ideally, I’d like to use the company laptop to read my own mail, access my NAS in my time off.
I fear that even if I configure containers on that laptop to run alpine + wireguard client + firefox, the traffic would still be decrypted. If so, could you explain how the wireguard handshake could be tampered with?
What about Tor in a container? Would that work or is that pointless as well?
Huge kudos if you also take the time to explain your answer.
EDIT: A lot of you suggested I use a personal device for checking mails. I will do that. Thanks for your answers!
You must log in or register to comment.
Don’t. Just fucking don’t. Keep your personal stuff off your work equipment and vice versa. I don’t know why people keep wanting to do this, because it only leads to trouble.
Adding on:
Anything you do with a company device brings liability to them, which is part of why you should keep things separate, and part of why they manage devices.
The best thing is to use a different device, period.
Since the company is lord and master over the device, in theory, they can see anything you’re doing.
Maybe not decrypting wireguard traffic in practice, but still see that you’re doing non-official things on the device that are probably not allowed. They might think you’re a whistleblower or a corporate spy or something.I have no idea where you work, but if they install a CA they’re probably have some kind of monitoring to see what programs are installed/running.
If the company CA is all you’re worried about, running a browser that uses its own CA list should be enough.
Yeah, I’ll use my own device, log on to the guest network and start Wireguard on my laptop. Seems a fair choice both for the company and myself.
That seems more sensible.
But they still can track some of the things you do (same with any untrusted wifi network):
- all data of http traffic (i.e. non-https)
- ip addresses you connect to
- hostnames you connect to (if SNI is not working correctly)
deleted by creator
Thank you for sharing this info. It’s very convincing and well argumented.
I won’t try anything else and will use my personal device.
I’m curious on how your systems would handle something like a guacamole instance running on a users home network? It’s pure http traffic afaict, but I’ve always been curious how it would be logged.
Depending on the set up, but there should be something that logs all network connections. So they can see the connection to the private IP, just can’t see what it was
deleted by creator
Personal stuff on personal devices.
Company stuff on company devices.
Never mix. I don’t even check my personal email on my work laptop.
If I need access to my home, it’s through an external connection like LTE.
This, but if you use an OS like Qubes then you can keep them separate while still using the same device
I think that with 802.1X you can’t do that unless you export the keys somehow.
You wouldn’t do this with a stranger’s device, so why insist you do it with your employer’s device? Just don’t.
If you have a workstation and want to use the same monitors/headsets/peripherals with both the company device and your personal device try one or two KVM switches.
Power toys and mouse without borders would like a word
Don’t. And beyond that if you use their WiFi, connect to a VPN. Best just use LTE.
What you are wanting to do is likely a misuse of corporate resources. If you are still unsure go talk to IT
Use a tails usb
Preferably just dont though
I’d its tails then its not an issue at all. Its using the same keyboard but for all intents and purposes, its a different machine
If it boots from USB, boot a different OS. But overall, preferrably use a different device.
If you want to use the same physical device just put Linux on a bootable USB stick and boot off that
Do not do this if it let’s you. Its a good way to get in trouble
I’ve done this in the past without apparent issue. Could you perhaps expand on where the risks arise here? My impression was that unless there is some independent hardware running code separate from the OS, then it would be OK?
Not an expert, but I believe that the company could detect that it was booted to another OS and you could have trouble at work for policy violation. But this process would likely be ‘safe’ from a personal privacy/security perspective.
Its likely a violation of company policy
Let’s assume it’s allowed. Obviously it’s untrusted hardware, but for widely issued corporate PCs, what’s the risk that there would be some hardware snooping going on if you controlled the OS?
The “snooping” is called a good security policy. Security should always come first.
Just run portable Firefox without the root cert?
Which browser you use won’t really matter. The company is using an SSL proxy and they’re not going to pass your traffic along and let you bypass it. You don’t really get a choice as the end user. You can accept their proxy cert one time by adding it to your browser store or you can accept it every time you try to visit a site. In either way they’re going to decrypt the traffic and re-encrypt it.
FWIW the SSL proxy should only impact asymmetric encryption that uses TLS. It shouldn’t impact symmetric crypto but they can still monitor everything you do by other means. They can watch you and they can block any traffic they desire. Chances are if they’re willing to go far enough to deploy an SSL proxy then they’re probably willing to fire you if you try to bypass it.
It’s good to know that they can’t bypass wireguard or Tor. I was a worried about that.
As others have suggests, I will probably use a separate device to check my mail. That seems the safest and fairest option both from the company and my perspective.
Protecting your traffic over the wire also doesn’t stop them from getting the data directly from the OS or program itself.
It’s their hardware, you’re just allowed to use it (and according to papers you signed when hired, likely only for work use).
My company uses a similar MITM technique on all our network traffic, but we have also used a number of other tools that don’t have the ability to snoop on the network traffic but can still get browsing data from user machines. Most browsers have “enterprise mode” features, or just store browsing history in a file that other programs can read.
We’ve also used systems that installed at the BIOS and/or bootloader level to allow us to track the location of and take certain remote actions on company hardware that was taken off the company network. If the device got an internet connection at all, it was still ours to control. Was very handy for people who tried to keep their laptop after they quit.
Technically they could use OCR on automatic screencaptures, which would bypass anything you could do. There’s a ton of “management” software that does automatic screen captures, or allows someone to look at an overview of desktops like a security guard looking at a bank of camera monitors. Usually that’s something schools use, but it is available for companies.
They could use a keylogger too.
The point is, you cannot control, or have any foolproof knowledge of, what they have installed on your work machine. That means that you cannot effectively work around or bypass it. If you absolutely need to, make a new “personal” email account to use for things like spotify or youtube on your work machine, and just use your damn phone for personal stuff.
I tried opening a browser in a Docker container and but couldn’t browse any site except google because it didn’t recognize the CA authority.
Often gets blocked. I tried on my work and you can’t use any other browser.
What I did is use a ssh tunnel and rdp over that. ssh and RDP are both build in to windows. VPNs often don’t work because some software needs to be installed.
The computer probably has local security tools (such as an edr) that spy on you any way.
You need to assume it is completely compromised.
But… assuming this isn’t in violation of your company computer usage policy (which it very much might be and can put you in trouble) you can install any VPN (avoid spyware shit) and a different browser (ideally something a bit obscure, like librewolf) and this will bypass the MiTM as the the device that does the MiTM would be either:
A) a network device that hijacks the HTTPS requests (VPN bypass this)
B) the browser used by the company
C) some other kind of software that atteches itself to all browsers via admin installed extensions (obscure browser might not be recognised by such software, be sure to check the installed extensions after letting the browser run for an hour)
And once you are done you can check the certificate chain in the browser to confirm.
Before I wrote this thread, I ran for a couple of minutes a browser from a docker container. I couldn’t browse any website because of the missing CompanyName CA certificate. So, I stopped because it was too freaky.
That makes sense, the MiTM was still going on but you browser was not configures with the company CA
deleted by creator
Either double-fist with a second laptop or install QubesOS on your laptop
