While I was asleep, apparently the site was hacked. Luckily, (big) part of the lemmy.world team is in US, and some early birds in EU also helped mitigate this.

As I am told, this was the issue:

  • There is an vulnerability which was exploited
  • Several people had their JWT cookies leaked, including at least one admin
  • Attackers started changing site settings and posting fake announcements etc

Our mitigations:

  • We removed the vulnerability
  • Deleted all comments and private messages that contained the exploit
  • Rotated JWT secret which invalidated all existing cookies

The vulnerability will be fixed by the Lemmy devs.

Details of the vulnerability are here

Many thanks for all that helped, and sorry for any inconvenience caused!

Update While we believe the admins accounts were what they were after, it could be that other users accounts were compromised. Your cookie could have been ‘stolen’ and the hacker could have had access to your account, creating posts and comments under your name, and accessing/changing your settings (which shows your e-mail).

For this, you would have had to be using lemmy.world at that time, and load a page that had the vulnerability in it.

    • DreadTowel@lemmy.world
      link
      fedilink
      arrow-up
      3
      ·
      1 year ago

      Yeah, I know. It is possible to implement it though. Posts and comments are cached in every federated server, the only thing you need is attribution to the author. Your key/certificate could be your identity.

      • 0x4E4F@sh.itjust.works
        link
        fedilink
        arrow-up
        2
        ·
        1 year ago

        It’s… slightly more complicated than that, but it should be possible.

        This was one of my early fears… that an instance might get offline, attacked, whatever, and that all your user data might be lost. It appears that a feature like this might be inevitable. Or at least a feature that would let a user to have at least 1 copy of his account on another instance.

        There is also the DB and storage growth to consider if something like this is implemeneted. If all instances hold all data from every other instance, than that might turn into a shit show pretty soon. You can get away with a certain number of users on only one server, but after a certain threshold, you have to scale… and that costs a lot of money, which is a no go for smaller instances and they will probably shut down if they’re forced to hold data for every user account in the fediverse.

        • DreadTowel@lemmy.world
          link
          fedilink
          arrow-up
          1
          ·
          1 year ago

          If I had to propose something, I’d suggest writing a spec and implementing an API for Lemmy servers that would allow one to submit signed actions, like posts, comments, likes, display name, etc. Then people will write clients that will allow generation of certificates to be used to sign those messages + add the ability to export/back up the key to be used by other apps too.

          If storage space ever becomes an issue, you could potentially shard communities. I’m more worried about the network traffic. I already suggested a routing algorithm that would spread the load between all federated instances and would scale like O(log(N)). There was some interest, but it would be a long term project. There’s still a lot of performance that can be gained by simple optimisations.

          • 0x4E4F@sh.itjust.works
            link
            fedilink
            arrow-up
            2
            ·
            edit-2
            1 year ago

            Yeah, there are more pressing issues at hand, a lot of bugs, a lot of new features that need to be implemented ASAP. Like user following, Lemmy is still lacking that.

      • steltek@lemm.ee
        link
        fedilink
        arrow-up
        1
        ·
        1 year ago

        Accounts aren’t just posts/comments/votes. You have the communities you’re following, settings, and other bits of private state (remembering what you’ve already read). You’d be in charge of preserving and lugging this data around.

        • DreadTowel@lemmy.world
          link
          fedilink
          arrow-up
          1
          ·
          1 year ago

          Either store that data yourself or rely on your servers to store it and distribute it, just like they do today.

      • kakes@sh.itjust.works
        link
        fedilink
        arrow-up
        2
        arrow-down
        1
        ·
        1 year ago

        I’ve thought about this, and had come to the conclusion that it would be too difficult for users to manage their private key. (Look at the flustercuck we’ve seen in cryptocurrency)

        I wonder though, if we could choose to either manage our own private key OR let an instance manage it for us. This way, users could choose to give full control to an instance (as it is now) or to keep control themselves for extra security.

        User data and public keys would still need to be stored (presumably on an instance server, though I wouldn’t be against separating this duty into a separate server which may or may not be hosted concurrently with an instance). But at least this way, we would have the option to keep control over our accounts even if the instance is hacked.

        • DreadTowel@lemmy.world
          link
          fedilink
          arrow-up
          1
          ·
          1 year ago

          If I had to propose something, I’d suggest writing a spec and implementing an API for Lemmy servers that would allow one to submit signed actions, like posts, comments, likes, display name, etc. Then people will write clients that will allow generation of certificates to be used to sign those messages + add the ability to export/back up the key to be used by other apps too.

          I think people who actually understood crypto used it correctly. It’s just that most people were there to speculate and gamble. And, tbh, I think crypto is here to stay.

          Yup, I’d support an option to store the key on a server. Or just use an account tied to a server, just like now. Currently, servers sign all messages to other instances with their key, so all messages are tied to the identity of the instance!

          • kakes@sh.itjust.works
            link
            fedilink
            arrow-up
            2
            ·
            1 year ago

            100%. Also, while we’re at it, I just want to throw it out there that the front-end should also be split out imo. Let people mix and match backend/user/frontend according to a central api specification, so the best options for each can rise to the top, rather than relying on a single monolithic codebase.