UnrealIRCd Forums

Hi,

I have the following kind of bot attack in my server. Everytime a user says some msg in the main channel, a nick (bot) will reply the same msg to the sender as notice msg in reverses and underline. Any regex or module for that? Really thank you in advance.

P.S. There is another similar attack when users who type !seen to search for users, they will get a reply msg from a bot (script) who send replies. Need help for that too.


Thank you all !

Underlines are ignored by spamfilters, and you could always reverse something that is spamfiltered and say that causing the bot to say it triggering that spamfilter and banning the bot.

As for the !seen response, thats a common script and IMHO users will hate to have their !seen scripts blocked.

Hi,

Thanks for the help. Ok so it means we can't spamfilter reverses and underline. Instead we should just check for the malicious bot to trigger its response and ban it? What if the bot comes again and again with different IPs and nicks? You know, we can't spamfilter its reply notices as it's using users own messages in the channel to reply. Not fixed messages.

Thats why I suggested you typing somthing YOURSELF in reverse from how youve spamfiltered it, causing the bot to reverse it making it normal again and then triggering the spamfilter.

Say for example you have "!killme" spamfiltered with action kill, and you say !killme in channel, the bot sees it then reverses it to "emllik!", how about you then type "emllik!" yourself so when the bot reverses it it becomes "!killme" thereby triggering the spamfilter.

That then saves you the trouble of typing a full /gline command yourself and letting ths spamfilters deal with it.

Just an example, of course youd need to add a spamfilter for !killme yourself.

Jobe1986 wrote:Thats why I suggested you typing somthing YOURSELF in reverse from how youve spamfiltered it, causing the bot to reverse it making it normal again and then triggering the spamfilter.

Say for example you have "!killme" spamfiltered with action kill, and you say !killme in channel, the bot sees it then reverses it to "emllik!", how about you then type "emllik!" yourself so when the bot reverses it it becomes "!killme" thereby triggering the spamfilter.

That then saves you the trouble of typing a full /gline command yourself and letting ths spamfilters deal with it.

Just an example, of course youd need to add a spamfilter for !killme yourself.

No no...what i meant was not reversing the words. It's reverses and underlines <<<< color codes. The bot is repeating whatever a user typed and replying to him in reverses and underlines.

Annihilator: You can simply type !killme. The bot will add the colorcode stuff, which spamfitler will ignore. Thus, spamfilter will see !killme, and nuke the bot. You, as an oper, won't be hit because opers can't trigger spamfilters.

Well... if the bot is doing it as a notice then it doesn't matter. The spamfilter just needs to be setup to gline the notice with whatever keyword he chooses (such as the !killme) and the spamfilter will only nuke people who send the notice saying "!killme."

Thanks Jason and SpaceDog. I get it now.