XMPP Service Operators - 2026-06-06

mightyBroccoli, sorc60@magicbroccoli.de

https://downloadable.pain.agency/file_share/019e9ba5-dd8c-75fb-87c4-1e15cd620525/66d78341-b640-4b2a-b95f-492ff54d73d5.png

(idk if that sent twice, a mod did mass moderations and caused my server to hang long enough to drop connections)

also interestingly, it seems to have somehow gotten around my muc_limits, or perhaps i just set them too high

> recently there has been a spamwave with usernicks matching "JEW <some number>" and the message "am israel hi" or something like that > > does anyone know which server(s) those accounts come from? stratself: just from random servers who have open registration, not much to do other than ban/rtbl and carry on ↺

some channels just need more moderators :)

are there modbots that subscribe to the blocklists, but doesnt need access to the MUC components?

i only saw RTBLs being mentioned in terms of modules so i'm not aware of anything else

> are there modbots that subscribe to the blocklists, but doesnt need access to the MUC components? yep stratself ↺

https://xmppbl.org/channels 2 at bottom , I think there are some more but I don't have them handy

Oh, the original repo for riddim is gone. Kris grabbed a copy and uploaded it to https://f-hub.org/XMPP/Riddim but it is unmaintained currently

thanks, i was about to ask for the repo

do you have a recommendation between the two?

they all seems to be unmaintained for a few years now

That's the bot on the page that moparisthebest pointed out. I shared that since the link to the bot on the page is dead

If you can go with the module, go with that. It is still maintained and works more reliably

thanks, i'll check out the modules anyhow. though it'd be nice to have abilities to mod rooms that i can't really selfhost

i have some semirelated questions: - are there lists other than xmppbl that i can subscribe to? - on xmpp, is it easy to check if a server is open registration (no rate limits, no captcha, no nothing just straight up an open door for automated spam)? if it could be measured, it could be one of the more proactive blocklists i think

JoinJabber has a rtbl (real time blocklist). As for easily checking if a server has ibr (in band registration), there is trying yourself but anything automated is less reliable, if I remember correctly ✏

> i have some semirelated questions: > > - are there lists other than xmppbl that i can subscribe to? > > - on xmpp, is it easy to check if a server is open registration (no rate limits, no captcha, no nothing just straight up an open door for automated spam)? if it could be measured, it could be one of the more proactive blocklists i think I'm not sure how you would check for no limits etc in an automated way. ↺

well in a good way that isn't DoS'ing them anyway

Personally I don't feel the rtbl adds much, and has had controversy around who was added to it. At the end of the day, just make sure you have moderators you trust and most spam is solved that way. Also during spam season, set your channel to moderated, and voice those you trust, after 24-48 hours, you can drop it and let people speak freely again.

In 99% of cases this works well

There's never been any controversy except for a mistake that was made one time (which also I'd say wasn't very controversial, just a mistake that was easily corrected)

But yeah if you're not getting attacked by the gore spammer or the isreal spammer or whoever it maybe won't help you

I do appreciate seeing a spam wave start in another channel and never make it to mine due to rtbl. But YMMV it's just a tool

> Personally I don't feel the rtbl adds much, and has had controversy around who was added to it. At the end of the day, just make sure you have moderators you trust and most spam is solved that way. > > Also during spam season, set your channel to moderated, and voice those you trust, after 24-48 hours, you can drop it and let people speak freely again. my two cents is that defense in depth is better than having your mods not sleep. Ideally I'd like some sort of script-based spam recognition i.e. force the channel into voice mode and ban those spam accounts when it recognize irregular joins+traffic patterns and have it run without being a module too, for good measure ↺

Always better to be a module IMO. But also I try to avoid going into "moderated mode" in a public channel

Maybe there should be a module that autobans someone if they join a room for a first time and immediately starts spamming

Which is the case for all of the israel spammers

automatically classifying it as spam is the hard part

Sunglocto (sunglocto.net), how would you classify spamming? I usually get caught in such spam filters

Seen as I type too fast and have an itchy return finger

> Maybe there should be a module that autobans someone if they join a room for a first time and immediately starts spamming I also mute non members who post at a certain speed. Which is similar ↺

> I also mute non members who post at a certain speed. Which is similar You would mute me? :( ↺

> Sunglocto (sunglocto.net), how would you classify spamming? I usually get caught in such spam filters someone saying "🇮🇱🇮🇱shalom" 500 times a minute is probably easy to detect ↺

hindsight being 20/20 yes, but what's the next spam wave gonna be?

same word repeated N times per minute?

or per seconds

> are there modbots that subscribe to the blocklists, but doesnt need access to the MUC components? theres creme [envs.net]'s adminbot if you want, though not sure why he doesnt just use the prosody module ↺

muc_ban_bot? that's not automated though, but it can ban same jid on all the watched MUCs

> same word repeated N times per minute? this one wasn't that though, instead a bunch of different symbols to make a flag ↺

oh, okay

> muc_ban_bot? that's not automated though, but it can ban same jid on all the watched MUCs idk if thats the one that subscrubes to banlists ↺

> this one wasn't that though, instead a bunch of different symbols to make a flag often it's the same word/phrase though ↺

> often it's the same word/phrase though it did seem like you could probably just take a hash of the message ↺

> idk if thats the one that subscrubes to banlists me neither, I used muc_ban_bot once when they first announced it ↺

s/used/tried/

https://git.envs.net/envs/muc_banbot

looks like there's a lot more stuff now

> looks like there's a lot more stuff now I think creme likes to vibecode so thats what you get i suppose ↺

regardless it'll be a neverending arms race, if many deploy something to ban repeated words they'll change them a little each time

i wish mod_muc_limits had a per participant setting