THIS IS A STATIC MIRROR OF USERSCRIPTS.ORG - LOGINS DO NOT WORK

Spam and malware

in Userscripts.org discussion
Subscribe to Spam and malware 3530 posts, 606 voices



znerp Scriptwright

Another forum spammer - https://userscripts-mirror.org/users/62849

 
znerp Scriptwright

This looks like another spam topic - https://userscripts-mirror.org/forums/4/topics/3095

 
znerp Scriptwright

Another topic spammer: https://userscripts-mirror.org/users/62961

 
atearwhofellnot User

https://userscripts-mirror.org/users/63456

 
Norman Rasmu... Scriptwright

https://userscripts-mirror.org/users/63842

 
Aquilax Scriptwright

Help!!! This forum is drowning in the spam!!!

 
Mikado Scriptwright

What about one new topic per user per day restriction?

 
LouCypher Scriptwright

What about captcha? They're not human aren't they?

What about one new topic per user per day restriction?

That'll be good too, only the problem is it's using multiple accounts
https://userscripts-mirror.org/users/64120
https://userscripts-mirror.org/users/64121
https://userscripts-mirror.org/users/64122
https://userscripts-mirror.org/users/64123
https://userscripts-mirror.org/users/64124
https://userscripts-mirror.org/users/64126
https://userscripts-mirror.org/users/64128
https://userscripts-mirror.org/users/64129
https://userscripts-mirror.org/users/64130

 
Jesse Andrews Admin

Unfortunately "topics per day" will probably result in people spamming other topics.

I'm thinking limit new users to 5 posts the first day, 10 the second, 15 the third, ...

 
dob Scriptwright

What's your position on Captchas, Jesse?

 
Jesse Andrews Admin

I'm doing recaptcha on signup, perhaps doing captcha during posting for the first few posts?

 
dob Scriptwright

Yeah, maybe you'll be captcha'd for your first two threads or so?

 
Aquilax Scriptwright

Spam topic
Real Estate Portal Script

 
znerp Scriptwright

All of these users beginning with "Doctor" seem to be spamming.

 
pwlin Scriptwright

Spammers:
https://userscripts-mirror.org/users/66605/comments
https://userscripts-mirror.org/users/64199/comments

 
znerp Scriptwright

spammer: https://userscripts-mirror.org/users/67483

 
ekbworldwide User

>> Yeah, maybe you'll be captcha'd for your first two threads or so?

I think that's an excellent idea - spammers are lazy and opportunistic. I bet some of them when faced with a couple of captachas will go away.

 
znerp Scriptwright

I don't know if anyone would agree, but this looks like spam to me;

https://userscripts-mirror.org/forums/2/topics/9234
user - https://userscripts-mirror.org/users/67886/posts

 
Aquilax Scriptwright

This also is a spammer: https://userscripts-mirror.org/users/67936/comments.
He has posted 17 messages in 8 minutes ...

 
znerp Scriptwright

spam commenter: https://userscripts-mirror.org/users/61793
comments: https://userscripts-mirror.org/users/61793/comments

 
znerp Scriptwright

spam forum user: https://userscripts-mirror.org/users/68753
topic: https://userscripts-mirror.org/forums/2/topics/9311

 
pwlin Scriptwright

Spammer:
https://userscripts-mirror.org/users/68752/comments

 
khopesh Scriptwright

more spam: https://userscripts-mirror.org/users/61894/comments

There's already a +/- system for comments, with a threshold of 0 so that spam will theoretically get pushed below the visible threshold (like [slashdot](http://slashdot.org/)). I think that's well executed.

To add to that idea, though this might be too computationally complex, a Bayesian algorithm, trained on posts that get pushed below zero versus posts that get pushed up or posts that are oft-viewed by signed-in users, might help create an index for admins to find spam before it is either reported or moderated down. A Bayesian detector could also be trained to find malicious code, but you'd have to hack at the tokenizing code to better fit javascript rather than email content (Bayesian algorithms are not limited to spam detection!).

Also possibly too complex: instead of using slashdot-style +1/-1 moderation, use a scale of 0-5 and average the number for the resulting score (like [kuro5hin](http://kuro5hin.org/)). Update: Karma and moderation is further discussed here.

Of course, a simple index of negative-scoring comments for admins (without using Bayes) would be a good start, too...

@Aquilax mentioned a frequency issue ... maybe put an aggressive rate limit on any account that hasn't yet posted ~10 comments? Something like a limit of one post per 30 seconds is a good overall rate limit, and a limit of one per five minutes would be good for new accounts (a bot who hits its limit usually doesn't try to determine when it can next post, so that would stop a bot after a single post, whereas a per-day limit would see several posts before the bot is abandoned). Perhaps instead of the more aggressive limit, a @Jesse's suggested captcha for new users' first few posts ... but don't forget that manually entered spam exists, too. The nice thing about rate limits is that you can apply them to an IP (or even IP block) rather than a user (this solves @LouCypher's workaround).

 
jerone Scriptwright

Scrap all spammer: https://userscripts-mirror.org/users/39307/scripts
A lot of scrap all: https://userscripts-mirror.org/users/36222/scripts & https://userscripts-mirror.org/users/41738/scripts
Don't know if this allowed; a lot of the same scripts: https://userscripts-mirror.org/users/53421/scripts & https://userscripts-mirror.org/users/7563/scripts

 
JoeSimmons Scriptwright

Script copiers (copied from me):
https://userscripts-mirror.org/users/62984/scripts
https://userscripts-mirror.org/users/67111/scripts