RBL like for Mastodon moderation and spam fighting

One of the tools to avoid email spam is Relay Block Lists (RBLs) and although it’s not perfect, it can be a great help to block the most obvious cases of spam/abuse.

The other day I was looking into the RBLs I am subscribed to for an email server I run and thought that maybe a similar feature could work for Mastodon.

An RBL like could be a good fit because it is not a fully centralised, you may subscribe and unsubscribe to as many RBLs as you like, like the current Mastodon Relays.

I am just sharing this here to know what your opinions are on this and what could be problematic with this approach.

There are several ways this could be implemented in terms of where the RBLs are server, either using external software (like the current Mastodon Relays) or internally and included in Mastodon. In my mind an internal solution would be great because each instance could be an RBL.

Example: my instance already has domain blocks and email blocks, if you subscribed to my instance RBL then you would also be applying the same domain and email blocks.

Obviously there should be an option to make your instance an RBL (make your blocks public) or not.

I don’t think this would bring a great overhead, as RBL update frequency could be very low (like once every 24 hours or something, no need for realtime).

This could be a great way to reduce moderation needs and also reduce registration spam as the problematic emails would rapidly reach instances.


1 Like

i think it would be a very bad idea to have this for domain blocks, especially given the significant problems with oversight, transparency and accountability shared blocklists have had on twitter.

email blocks seem less problematic but i think there are probably better ways to handle spam.

In my experience, instances that get a lot of spam are already the bigger instances that should have moderation in place to handle it (i’m including masto.host instances in this bucket since they all share the same infrastructure). I don’t see a lot of small instances getting spam, but maybe i’m not looking in the right places.

I think the most useful thing we can do currently is surface the existing email/IP information in a more aggregated and mod-visible way, so it’s easier to figure out what the “common thread” is between any given spam attack. Lots of spam attacks use very common email domains, so the email blocklist isn’t effective against them. Most spam attacks require creativity or contextual information to identify, and they’re not amenable to a one-size fits all blocklist.


This has nothing to do with any specific need I have for Masto.host, it was just an idea I had for Mastodon.

On my instance, I manually copy paste domain blocks from a couple of instances I trust and it hasn’t been a problem but I am a very atypical admin, I know :slight_smile: probably most admins do it differently.

Thanks for the feedback.

This is a really good idea and could defiantly work with each server opting in to offer up their block list for others to subscribe to. There’s always going to be larger server and smaller servers, and as a smaller hobbies server it is unfeasible to stay on top of all the new baddies rearing their heads in the fediverse.

Some changes that must be included however would be a generalized category system for domain blocks and a note section that would be included with those rbl. That way admins that subscribe could make informed decisions of a servers block list aligns with their own CoC.
This isn’t an issue that is going to go away and has had to have been solved before when we went from arpnet to a larger networks, and this has been more effective then not implementing anything at all for email.

1 Like

I agree, this was just a general idea for discussion and there are many things I didn’t consider.

I also think that domain blocks should be manually approved or something like that, the admins get a notification similar to the current report and they can act how they see fit.

There was also a suggestion from @kev@fosstodon.org (thread) to have a checkbox in the domain block form to report that block as spam/RBL to any instance subscribed to that instance RBL. This way the reporting could be opt-in on a domain by domain situation. I like that idea, too.