Forum:Combating spammers

Forums: Index > Village Dump > Combating spammers

Note: This topic has been unedited for 2240 days. It is considered archived - the discussion is over. Do not add to unless it really needs a response.

Click here to edit this page if it's archived.

Hi, I've noticed that you suffer from massive spamming, as also seen in Illogicopedia. I have an idea to deal with it, and thus am here to post my proposal.

I know you guys have AbuseFilter, and I think you should use it. In fact, in Usopedia, where I'm an active admin, we are able to prevent spammers' posts with 100% accuracy using the filter (though it was easy, because what I needed to do was to block non-Japanese posts, which, I know, won't work that way here).

I should also tell you that Illogicopedia is discussing the same issue. So why not follow them?

Although it's not as easy as Usopedia for English Uncyclopedia to deal with spammers, I still have some specific ideas:

1. Disallow creation of pages without categories and/or templates that add categories (as being an encyclopedia, categories should be added, but spammers generally won't). - This might also help to disallow creating pages in just a few words or so.
2. Use the way proposed in Illogicopedia at the link above, which is to disallow continuous <br> tags.

I guess there can be other kinds of useful filters, and I would like to discuss on this matter, including the proposals for new ideas.

Thanks for reading.--The Pioneer (Saloon/Shootouts) 05:32, 5 July 2018 (UTC)modded.--The Pioneer 08:57, 5 July 2018 (UTC)

It's getting to be a huge problem, and it's almost humorous the shit they come up with. First it was Ralph Lauren, and since then it's been Nike, poker, rooftops, and all kinds of fuckery. I may be one of the most active admins, but I'm not the most savvy to these kinds of things. I've only been an admin since April. I wonder what Leverage or Kippy or Frosty could say about this; they're more senior than I am. PF4Eva, the President of Imagination Vote for me My tax returns 08:16, 5 July 2018 (UTC)

Hmm. If so, since Kippy and Leverage are listed here, I'll ask them to join this forum in their talk page (hope they won't regard me as another form of spam, because, you know, posting almost same sentences to several admins is an action that make them cautious).--The Pioneer 09:28, 5 July 2018 (UTC)

Sounds like a good idea, but we'd lose the ability to leave snarky remarks in the banlogs.... :( -- Sir Mhaille  (talk to me)

You don't need to worry about it. Not all vandals are spammers, as you know, and even the best AbuseFilter cannot prevent human vandals completely, which means you should still ban those bypass the filter. Besides, the AbuseFilter has no machine learning function, and thus never be capable of triggering technological singularity. (Or you know, you can just use Socky to generate prank ban logs :p)--The Pioneer 14:52, 6 July 2018 (UTC)

Sounds good. As far as I know, the people to prod and poke about this are user:Bizzeebeever or User:Lyrithya. The chatroom is sometimes the place where things like this get done. Leverage (talk) 18:20, 5 July 2018 (UTC).

OK, I guess I should poke them, too (though it seems to me that they have been inactive for several months). Besides, I need to talk to Lyrithya (or Athyria on Illogicopedia), because she seems to be the only active sysadmin there, who can edit interwiki links to officially make Usopedia a sister wiki. I'm not sure whether I can always be active on IRC, because I live in JST (UTC+9) time zone and most of you guys probably live in different time zones.--The Pioneer 05:44, 6 July 2018 (UTC)

I have nothing to offer here, as I'm pro-spam. -- Kippy the Elf Talk Works ☃ 04:42, Jul. 12, 2018

I have the solution: death sentence! The SRU, Spam Removal Unit, weeeeeeeeee... We have a device stolen to Mossad: thanks to it, we can transform the spammer in a frenchie frie with a single ZOT! Once he saves his promotional edit, he has condemned himself t5o an electric death, thanks to the geolocalization and to an electric impulse. His PC, tablet, smartphone or any device will be, instantaneously, a super fast electric chair, singing "Remove Spam From Premises!" If a second spammer from the same town dares to spam here, 20 driverless stealths will bomb dat town and erase it from Earth. So, the solution is: death sentence to spammers, and death sentence to pro-spammers and spammer supporters too. Our device has a holographic firing squad ready to fire. And ready to squad. Spammer's region, in a diameter of 150 km will be farted with a farting stealth. That region will smell of farts for 200 years. Our device, the SPAM REMOVER FROM PREMISES™, is ready to de-spam. Mwahahahahahahahah! --Spam Removal Unit 17:17, 12 July 2018 (UTC)

Mwahahahahah! Farts and frenchie fries for all! Mwahahahahah! --Spam Removal Unit 17:18, 12 July 2018 (UTC)

Mwahahahahah! --Spam Removal Unit 17:18, 12 July 2018 (UTC)

Oh yes. To use REMOVER™ you need some accordion lessons. The REMOVER™ is an accordion-shaped device that is activated by playing dummy Serbian songs. Once activated is unstoppable! Spammers are warned: surrender or become a crisp! Xaxaxa! Hue hue hue! --Spam Removal Unit 17:33, 12 July 2018 (UTC)

Yes, I answer to myself. Have I said dat I'm retarted and a fart smeller? --Spam Removal Unit 17:33, 12 July 2018 (UTC)

Nonsense. We don't want a Mossad fantasy, nor do we need to kill spammers. We just want to shut them up.--The Pioneer 02:33, 13 July 2018 (UTC)

Actually, we is of stolen to Mossad. But the REMOVER™ is not of killing machine. The electric impulse transforms the spammer, instantaneously, in a potato, and fries it one microsecond later. An is of salting him. A ready salted bunch of chips. And now, with this extraordinary offer, you can remove spam from premises with only 19.99 Serbian dinars. Remember, with only RSD 19.99 you'll is of receive the REMOVER™ and a complete set of Miracle Blade III. The perfect series of knives recommended by Chef Tony (#1 aliv in Serbia) and Tupac (#1 aliv in Serbia). Remember, call us, and remove spam from premises. Ohhhh yeah. --Spam Removal Unit 03:39, 13 July 2018 (UTC)

Abusefilter

Can't admins just... add/edit filters? I'm asking because I have no idea how we configured this. -— Lyrithya ༆ 11:56, 18 July 2018 (UTC)

I don't know how to use them. Who can run maintenance scripts? I found this which might get rid of all teh fake accounts we currently have. https://www.mediawiki.org/wiki/Manual:RemoveUnusedAccounts.php In terms of Spam filters, maybe the most obvious would be not to allow new users to create a page till they have made one edit elsewhere? I am assuming the spam is all done by bots right? Leverage (talk) 13:35, 18 July 2018 (UTC)

I do know how to edit the AbuseFilter, and can tell you that you can learn how by reading this, or I can share some tech using e-mails (I don't want to make it public, because I know that even English wikis aren't completely free of Japanese LTAs). You just need to go to Special:AbuseFilter (and no need to change configs), by the way. As for the script mentioned above, I know other uncyclopedias hosted by Carlb use it, but I'm not sure whether it's working all right.--The Pioneer 16:36, 20 July 2018 (UTC)

Lyrithya told me some tips on e-mail, and I agree with it, so will share them (in my words) to you.

1. How to avoid mistakes on legitimate contributions (especially that of new users: you can simply make an exception for "autoconfirmed" or any other groups)
2. Display a proper warning message through MediaWiki:abusefilter-warning-** (showing meaningful way to resolve the problem (ex. leading users to add categories on detection filter of pages without categories, etc.))
3. Proper way of actions (how they are affected / can they appeal if they were blocked/disallowed by mistake, etc.)

You can test a filter before implementing it, and I recommend you to do that beforehand.--The Pioneer 17:03, 20 July 2018 (UTC)modded brackets.--The Pioneer 17:10, 20 July 2018 (UTC)

Definitely test (run on recentchanges or such), and if you're still/really unsure about something, you can run a long-term test by instead tagging instead of blocking edits, or such... -— Lyrithya ༆ 03:59, 2 August 2018 (UTC)

---

It shouldn't be archived, as no one has drawn any conclusions, and no actions have been taken so far.--The Pioneer 23:54, 30 July 2018 (UTC)

Here are my ideas

Maybe someone (I'm not quite that tech savvy) can add the following filters:

Edit count <1: new line contains any of the following words: (This shouldn't effect autoconfirmed or admins)

• "Ralph Lauren"
• "Nike"
• "DSquared"
• "poker"
• "Rooftop" or "rooftops"
• "turd" or "turds" (You wouldn't believe the shit I've seen. Pardon the pun.)
• "plos into bannings me" (I already took the liberty)
• Non-English characters
• <b> or </b>
• "brainwash(ed/ing)"
• "liberal(s)"
• "conservative(s)"
• "Democrat(s)"
• "Republican(s)"
• "bias" or "biased"
• "fake news" (We've been getting some right-wing lunatic conspiracy theory spam)
• "Associated Press"
• "Reuters" (People have been spamming news articles verbatim)
• Anything to do with the Top (#) YouTube Alternatives
• "Lydia Smith" (unless somebody named Lydia Smith ever becomes famous; I had to protect Lydia Smith from creation due to this problem)
• Prevent them from creating articles with long titles made up of random words with poor English (e.g. "17 How to Get Free Nike Shoes from Ralph Lauren while playing poker ")
• Pages created by new users must have links, if not categories.

That should take care of most of these guys. Anyone have the techy know-how to do this? PF4Eva, the President of Imagination Vote for me My tax returns 07:28, 2 August 2018 (UTC)

On second thought, I would make exceptions for UnNews, due to a certain orange politician. PF4Eva, the President of Imagination Vote for me My tax returns 07:36, 2 August 2018 (UTC)

Well, probably the most difficult part is non-English characters. For the other part, I certainly know how to work on it :)--The Pioneer 08:02, 3 August 2018 (UTC)

I created a filter for Ralph Lauren. I don't know if I did it right, but I hope it helps. PF4Eva, the President of Imagination Vote for me My tax returns 10:43, 7 August 2018 (UTC)

I disabled it until further review, as it seems to be affecting Special:UserLogin. Not sure if that's good or bad. Now I can rest easy and worry about it when I wake up again. PF4Eva, the President of Imagination Vote for me My tax returns 16:15, 7 August 2018 (UTC)

I would not use regex when possible, because it can sometimes act in a weird way. I would write the filter like:... I'll just e-mail you for more info. Filter info shouldn't be public in general.--The Pioneer 07:19, 8 August 2018 (UTC)

Thanks, man. Rewritten from the ground up. Hope this works. PF4Eva, the President of Imagination Vote for me My tax returns 11:33, 8 August 2018 (UTC)

My pleasure.--The Pioneer 13:15, 8 August 2018 (UTC)

And yeah, I would choose to use or (|) for the final two conditions, if I use both of them (it'll be like: & (old_size ~ | !contains_any ~ ).--The Pioneer 13:20, 8 August 2018 (UTC)

Let me know once you find the Swiss knife anti-spam solution. Until that, AbuseFilter is probably the best bet, just like how plain talk pages are the best way of implementing discussion inside MediaWiki (in other words: it really isn't, but...).

What makes AbuseFilter great for blocking spam and other unwanted behavior is that it's extremely versatile. What makes talk pages great is that it's basically a blank sheet of paper — totally unstructured. What makes AbuseFilter sucky is that it's versatile: you need to implement the filters, AF does not ship with any filters preinstalled and indeed, as mentioned earlier, AF lacks machine learning capabilities.

One thing to consider would be implementing QuestyCaptcha instead of the current FancyCaptcha. FancyCaptcha is the traditional "distorted word" CAPTCHA, which can sometimes be hard for humans to decipher but is rather easy for spambots to break, sadly — slightly better than the default SimpleCaptcha where users are asked to solve a simple mathematical equation, but only slightly. However, some things to consider if and (hopefully) when implementing QuestyCaptcha:

• there still isn't an on-wiki interface (like what Special:Interwiki is for the interwiki table) for viewing, modifying and deleting questions (and their answers) → sysadmins need to manually add the questions & answers to the configuration file
• coming up with good questions is hard. There needs to be a sufficient pool of questions and the questions should require some thinking, but they should definitely not be too complex (e.g. this one programming-related wiki asks users to run a shell command and paste the output (current timestamp) into the signup page) because that'll affect legitimate users negatively by driving them, too, away.
• but it is pretty damn effective! If you browse Brickipedia's new user log, you can pretty much see when QuestyCaptcha was implemented — virtually no spam registrations after that.

-- Jack Phoenix, professional killer admin (Whine?) • (Wikia ♥ ads) 22:53, 8 August 2018 (UTC)

I just tweaked the Plos into bannings me filter to include the somehow-even-more-bizarre LEET version, and the words "turd" and "turds." Where the hell are these spammers coming from? PF4Eva, the President of Imagination Vote for me My tax returns 03:14, 12 August 2018 (UTC)

I went one further and applied the filter to ALL article spaces. If that doesn't stop Mr. Turd, I don't know what will. PF4Eva, the President of Imagination Vote for me My tax returns 20:57, 12 August 2018 (UTC)

I hope I did this right, but I've added a condition that prevents potential spammers from having a username with an admin's username in it. (e.g. I wanna be huffbanned by PF4Eva; I wanna lick (admin's name)'s stinky turds). I hope this works/helps. The PIBM and Ralph Lauren seem to be working just fine. (Bad syntax, so forget it.) PF4Eva, the President of Imagination Vote for me My tax returns 01:17, 16 August 2018 (UTC)

It's been a crazy day. Lost Labyrinth decided to block Spam Removal Unit for using multiple IP addresses. And from his ashes have rose Spam Unit Removal Reloaded (not sure if he's legit) and, curiously, Spam Allowing Unit. After I banned Allowing, someone appropriately named Lazarus went on a spamming spree. It's been a mess to clean up. We need all the help we're gonna get. The original SRU was a step in the right direction, IMO. This has clearly gotten out of hand. PF4Eva, the President of Imagination Vote for me My tax returns 02:17, 19 August 2018 (UTC)

Things are getting weird (but you know, there are indeed some users who do such things; a vandal pretending to be helpful, mostly to gain advanced permissions)... Btw, I think Lazarus is one of the vandals from commons or UnMeta; I saw that kind of edit summaries written in capital letters there.--The Pioneer 17:14, 26 August 2018 (UTC)

Spammers in Simple Uncy

I apologize for the intrusion but as sysop in UnMeta/UnCommons I should report cases of spam attacks that happened in Simple Uncy where it is inactive. I would like to hear ideas and that some users may know about the situation. Generally, attacks on UnMeta/UnCommons are reversed with some delay but it is somewhat unsatisfactory to see the little Uncy abandoned at the mercy of vandals and spammers. Rhubella Marie, the rat sockpreppie preppiedits 13:11, 19 August 2018 (UTC)

Solution: Give trustworthy people rollbacks, and let them have at them. Better yet, use a "rollback all" tool to nuke vandal mass edits. -- Simsilikesims(♀UN) Talk here. 04:56, 16 October 2018 (UTC)

Solution#2: Protect pages to admin editing only, or autoconfirmed and admin editing only. Solution#3: Make massive use of the Abuse Filter.

PS: Sorry this response took so long: you may reach me quicker on Facebook. Yes, I have an account on Facebook under Simsilikesims now. I use it for gaming mostly, but in an emergency situation you can message me there. (I also have two other accounts on Facebook BTW, so I may or may not see it the same day, but I will see it the same week.) -- Simsilikesims(♀UN) Talk here. 05:01, 16 October 2018 (UTC)

Some (too much Draconian?) ideas

Well, this nite and yesterday evening 8 new IPs created their crappy spam. We need to wrap this up... Some of my ideas should be:

• A massive rangeblock related to any of this IP
• A temporary editing block for IPs
• We should consider the possibility to do as on English Wikipedia: new articles should be created by autoconfirmed users

Too much Draconian? But the rangeblock should be considered as soon as possible. Btw, why the "user creation log" is so full of new people every day? Lol, we have more users than all Wikmiedia :-D --Spam Removal Unit 06:04, 7 October 2018 (UTC)

Spam 3Yfq9Vo38E 07:30, 15 October 2018 (UTC)

Spoom spom spem. Kawang! Badong! --Spam Removal Unit 23:08, 16 October 2018 (UTC)