You can make it 100% effective by detecting the first spam message, ban the user, remove the message. There would be no way around this.

How did you solve the problem of banning the user before messages are sent? Is there a unique identifier or is the spammer bot that dumb that it joins with name "nuclear launch detected"? :D

Edit: Saw you mentioned the link but this cant really be a safe method. Banned any innocent folks yet?

Detecting dubious name changes maybe is an option too.

Speaking for bwcl specifically, banning innocent people by error 5% of the time (or less) is worth it and has no consequence beyond a tiny delay, clan leaders can contact admins directly outside the server and if a clan member gets banned by mistake their leader can let us know and we d fix it. So it works well for our specific use case.
That being said,I understand that it might be more problematic for cpl for instance

MeSaber

Sweden1235 Posts

August 03 2022 08:33 GMT

#82

On August 03 2022 02:44 Excalibur wrote:

Show nested quote +

The detect/ban/delete is already all working, but as not-the-guy-who-made-the-bot, I wouldn't want to speak as if I understand from a technical standpoint how it functions. 95%+ was a guesstimate to be fair to edge cases, but every server Defender is on has reported it working and catching problems before they start. Still, knowing technology, I don't like to claim 100% efficacy especially when both sides are free to make updates/changes daily.

In my layman's understanding, we have several types of detection involving user details, behavior, and messaging. (Edited at Nite's request.) Anything further would have to come from Nites rather than myself, and are not details I'd want to discuss in public view for obvious reasons. This is part of why the GitHub is not fully public as well.

We have had maybe one instance of a false-positive on SEN that I know of in the months we've been running Defender, and that was solved/patched same-day.

We also support Cross-Banning so once one Defender server detects a harmful user, that ban is propagated across all Defenders.

Its my opinion that Defender is the best way forward for anyone getting these targeted attacks.

Sounds great. If its been running for months how come zzzeros BSL discord hasnt gotten it yet. I left it because it was annoying as hell with the spam. I also dont add much to the discord so it was for the best anyhow.

Still BSL should in that case definitely have it.

Kudos for the bot maker for being awesome. I would definitely want to digest the source code but im ok with not knowing if you want it hidden from the spammer

Excalibur

United States59 Posts

August 03 2022 12:56 GMT

#83

On August 03 2022 17:33 MeSaber wrote:

Show nested quote +

You'd have to ask the Discord owner. If they have it installed but it isn't working they may have implemented something incorrectly. If they need assistance they can reach out to Nites or ask on SEN disc.

Nambona890

Australia32 Posts

August 19 2022 10:03 GMT

#84

Fucker is even willing to waste phone numbers now. He's hit at least one server that requires your Discord to have a verified phone number. Jesus Christ.

MeSaber

Sweden1235 Posts

August 19 2022 10:33 GMT

#85

Well 2-factor auth isnt built to protect a server but rather the user of the account from breach.

https://en.wikipedia.org/wiki/Multi-factor_authentication#Mobile_phone-based_authentication

Kinda naive to think this would work against someone who is this dedicated.

iPlaY.NettleS

Australia4383 Posts

August 20 2022 06:34 GMT

#86

I second what was said before about the star-defender bot.Moataz alts have already been autobanned five times since it was installed a month ago in the FFA discord, with all their messages auto removed.Great stuff.

Leonix

161 Posts

August 21 2022 03:17 GMT

#87

Can you keep Mozart just in peace, so he can focus on writing Music. It is getting rather dull...

chrisolo

Germany2609 Posts

September 16 2022 10:11 GMT

#88

To anyone still not convinced that this is Moataz. I think this is pretty much admitting to it. He used this Gif with his spambots. Sure it still is not proof, but if every indication leads to him, it is really likely that he is responsible for this.

I am just glad StarDefender is protecting most Discord servers by now and it seems like Discord added a feature that allows to prevent such attacks with their AutoMod feature.

But I still wanted to share this, to show that it is basically confirmed that Moataz is the perpetrator of the Discord attacks.