I prefer to pass secrets between programs through standard input

(utcc.utoronto.ca)

39 points | by ingve 2 hours ago

7 comments

  • kevin_thibedeau 2 hours ago
    > Unfortunately you're using a browser (or client library) that my anti-crawler precautions consider suspicious because it's sending inconsistent values for Sec-CH-UA-* HTTP request headers...

    The world doesn't exclusively use Chrome. Nice to see even the nerds are contributing to the closed web.

    • Alex-Programs 1 hour ago
      It's also moaning about me coming from a datacentre IP (proxy) with some vague complaints about load introduced by AI crawlers. I think this guy treats "protecting" his site as a hobby.
    • edwcross 1 hour ago
      I'm using Firefox and didn't see that message.
      • swiftcoder 1 hour ago
        Nor on Safari. I wonder what exotic browser the parent is using?
        • ErroneousBosh 1 hour ago
          Doesn't appear to be Firefox, Chrome, Chromium, Edge, or Falkon on Linux, doesn't appear to be Falkon on Haiku.

          I also wonder what they're using and where can I get some so I can break stuff too?

    • mhitza 1 hour ago
      Also site is not accessible via Mullvad VPN.
      • figmert 1 hour ago
        I am on Mullvad (at the router), and I am able to connect.
        • mhitza 56 minutes ago
          Checks out, it was my preferred exit node.
      • zxcvasd 1 hour ago
        [dead]
    • efilife 1 hour ago
      I am on ungoogled chromium and I see this
    • zxcvasd 1 hour ago
      [dead]
  • azornathogron 1 hour ago
    For one of my projects my server needs a private key, and it reads this from a file descriptor on startup and then closes the fd. The fd is set up by the systemd unit, which is also configured to restrict filesystem access for the server. So the server reads a key from a file that is never visible in its mount namespace.
    • computerfriend 43 minutes ago
      I do something similar with LoadCredential and it is quite amazing, especially when you want to run the application as a dynamic user.
  • juancn 1 hour ago
    I used to do that, I had a sort of IDE that launched a local server, bound to localhost.

    The launching process would send a random password through stdin to the child after launch, and the child would use that to authenticate the further RPC calls.

    It's surprisingly hard to intercept a process' stdin stream.

  • stale-labs 1 hour ago
    The main practical win is that cmd args show up in `ps aux` for anyone on the system to see, whereas stdin keeps it off that list.

    re: the /proc concerns - true, but if someones got same-user access to read your /proc/pid/fd, they can probably ptrace you or read process memory anyway. stdin is more about basic hygiene than stopping sophisticated attackers.

    tbh for anything actually sensitive I've been leaning toward tmpfs files with strict perms, or using something like vault/age. stdin is a nice middle ground tho for quick scripts.

    • reliefcrew 37 minutes ago
      > The main practical win is that cmd args show up in `ps aux` for anyone on the system to see, whereas stdin keeps it off that list.

      For those interested, re-mounting /proc with hidepid can prevent this:

          `mount -o remount,rw,hidepid=2 /proc`
  • Dwedit 2 hours ago
    I haven't actually tested this, but aren't the input and output handles exposed on /proc/? What's stopping another process from seeing everything?
    • trashb 1 hour ago
      Yes pipes are exposed /proc/$pid/fd/$thePipeFd with user permissions [0].

      Additionally command line parameters are always readable /proc/$YOUR_PROCESS_PID/cmdline [1]

      There are workarounds but it's fragile. You may accept the risks and in that case it can work for you but I wouldn't recommend it for "general security". Seems it wouldn't be considered secure if everyone did it this way, therefore is it security through obscurity?

      [0] https://unix.stackexchange.com/questions/156859/is-the-data-...

      [1] https://stackoverflow.com/questions/3830823/hiding-secret-fr...

    • Lex-2008 1 hour ago
      not a Linux expert, but I believe that at the very least it's time sensitive: after consumer process reads it, it's gone from the pipe. Unlike env vars and cli argument that stay there.
    • Tajnymag 1 hour ago
      I guess the kernel is stopping that. I don't think permission wise you'd have the privileges to read someone else's stdin/out.
  • pvtmert 1 hour ago
    Interesting approach. I like Docker/Kubernetes way of secret mounts where you can limit user/group permissions too.

    Meanwhile, I was an avid user of the echo secret | ssh consume approach, specifically for the kerberos authentication.

    In my workflow, I saved the kerberos password to the macOS keychain, where kinit --use-keychain authenticated me seamlessly. However this wasn't the case for remote machines.

    Therefore, I have implemented a quick script that is essentially

        security find-generic-password -a "kerberos" -s "kerberos-password" -w | ssh user@host kinit user@REALM
    
    Which served me really good for the last 4~years.
  • blibble 1 hour ago
    linux has a key api that works pretty well

    man keyctl