Hi all, we are hiring a remote worker and will be supplying a laptop to them. The laptop will be running a Debian variant of Linux on it.

We are a small shop and this is the first time we have entrusted somebody outside of our small pool of trusted employees.

We have sensitive client data on the laptop that they need to access for their day-to-day work.

However, if something goes wrong, and they do the wrong thing, we want to be able to send out some kind of command or similar, that will completely lock, block, or wipe the sensitive data.

We don’t want any form of spying or tracking. We are not interested in seeing how they use the computer, or any of the logs. We just want to be able to delete that data, or block access, if they don’t return the laptop when they leave, or if they steal the laptop, or if they do the wrong thing.

What systems are in place in the world of Linux that could do this?

Any advice or suggestions are greatly appreciated? Thank you.

  • notabot@lemm.ee
    link
    fedilink
    arrow-up
    40
    arrow-down
    1
    ·
    30 days ago

    There is a fundamental issue with this approach: the rogue employee has already copied the data to a USB drive by the time you try to wipe it.

    If the data is confidential, you either need to set up standard disk encryption and trust the employee, or not let them access it in a way it can be bulk copied. For instance, might it be possible for them to use a webapp that you control access to or a remote desktop type setup?

    • dan@upvote.au
      link
      fedilink
      arrow-up
      7
      ·
      edit-2
      29 days ago

      A lot of employers (at least the larger ones) block USB drives and have software to monitor for data exfiltration - monitoring where files are copied to, usage of copy/paste in browsers, etc.

      • notabot@lemm.ee
        link
        fedilink
        arrow-up
        2
        arrow-down
        1
        ·
        29 days ago

        I agree that you should assume you’re being monitored, but, while that helps against malware type exfiltration, it does little to stop someone who is determined to exfiltrate the data as there are a myriad of ways to do so that aren’t possible to monitor, such as simply taking a video of the screen whilst displaying the information.

        Ultimately, the company has to trust the employee or not give them access to the sensitive data, there’s no middle ground.

        • PowerCrazy@lemmy.ml
          link
          fedilink
          English
          arrow-up
          2
          ·
          29 days ago

          This is fundamentally true. However it is possible to limit the bandwidth of data the employee can exfiltrate.

          Assuming a privileged employee suddenly becomes a bad actor. Private-keys/certs are compromised, any kind of shared password/login is compromised.

          In my case I have a legit access to my company’s web-certs as well as service account ssh-key’s, etc. If I were determined to undermine my company, I could absolutely get access to our HSM-stored software signing keys too. Or more accurately I’d be able to use that key to compile and sign an arbitrary binary at least once.

          But I couldn’t for example download our entire customer database, I could get a specific record, I could maybe social engineer access to all the records of a specific customer, but there is no way I’d be able to extract all of our customers via an analog loophole or any standard way. The data set is too big.

          I also wouldn’t be able to download our companies software source code in it’s entirety. Obviously I could intelligently pick a few key modules etc, but the whole thing would be impossible.

          And this is what you are trying to limit. If you trust your employees (some you have to), you can’t stop them from copying the keys to the kingdom, but you can limit the damage that they can do, and also ensure they can’t copy ALL the crown jewels.

          • notabot@lemm.ee
            link
            fedilink
            arrow-up
            2
            arrow-down
            1
            ·
            29 days ago

            This is fundamentally true. However it is possible to limit the bandwidth of data the employee can exfiltrate.

            That’s fair, but the OP was talking about having the sensitive data directly on the laptop, which rather limits your ability to control access to it, and why I was suggesting keeping the data on a server and providing access that way, so limits can be put in place.

            Assuming a privileged employee suddenly becomes a bad actor.

            Your threat model probably isn’t the employee who suddenly goes rogue and tries to grab everything, but the one who spends and extended period of time, carefully, extracting key data. As you, the former can be be mitigated against, but the latter is very much harder to thwart.

            But I couldn’t for example download our entire customer database, I could get a specific record, I could maybe social engineer access to all the records of a specific customer, but there is no way I’d be able to extract all of our customers via an analog loophole or any standard way. The data set is too big.

            That’s well set up, but, whilst your competitor would love the whole database, what they’re really interested in is the contact details and contract information for maybe your largest three customers. When the dataset to extract is small enough, even quite advanced rate limiting can’t really help much. Making sure no one person has access to all of the most valuable data is a good start, but can present practical problems.

            And this is what you are trying to limit. If you trust your employees (some you have to), you can’t stop them from copying the keys to the kingdom, but you can limit the damage that they can do, and also ensure they can’t copy ALL the crown jewels.

            I think we’re basically saying the same thing. The OP talked about putting all the sensitive information on the employee’s laptop, and that’s what I was trying to steer them away from.

            In the past I’ve been asked if we can provide our developers access to pull the full source tree, but not copy it anywhere, and, to quote Charles Babbage, “I am not able rightly to apprehend the kind of confusion of ideas that could provoke such a question.”

            I enjoy the security side of sysadmin work, but I find it rather depressing, as all you can hope to do is lose slowly enough that it’s not worth attacking you.