Feminist SF Wiki:Wiki Spam

The first 4 entries on this page were copied from the Talk:Administrator Discussion page.

I am spurred to create a page for people (self-defined administrators -- yes, you too, whoever is reading this) to discuss issues relating to running the wiki. The spur is those long numeric strings that periodically show up on brand new pages that never existed before, or, lately, as additions to preexisting pages. What's up with those? They don't seem to be link-spam because there is no link. So I wonder if they are created by some searchbot or other program running wild, by the mediawiki software, or what? Does anybody know? LQ 07:01, 26 June 2006 (PDT)

Well, a search on the internet didn't turn up anything definitive. But some people theorize it's a "proof of concept" spam to see if the site is being patrolled. If we don't remove the weird edits, the real attack begins. Some pages I came across indicated that there are RBL filters that can be installed for the Wikimedia engine. Maybe we should look into that? --Therem 19:18, 26 June 2006 (PDT)

One geek friend believes them to be screwed-up spambots. Deleting quickly is a good thing. Looking for spam-filters would be great! Janice are you going to do that? LQ 14:04, 27 June 2006 (PDT)

Well, I have turned on the "sorbs" open proxy blocking option in the config file. I also added a page per the instructions at http://meta.wikimedia.org/wiki/Wiki_Spam. I am porting this page content over there now. --Therem 17:28, 27 June 2006 (PDT)

We're getting even more than ever! I'm not sure the "sorbs" has helped at all. (Or maybe it would have been worse without.) LQ 04:34, 30 June 2006 (PDT)

I note that a lot of the spam entries are intended to escape visible notice by including a < style = "display:none;" > tag -- is there a way for us to include entries that have that code on a watch list? -- LQ

And wiki responds -- Per the Wikimedia Anti-spam Features I added a $wgSpamRegex line. See also http://www.mediawiki.org/wiki/Manual:$wgSpamRegex for more help. (And I got dinged by it right away even editing the code that already existed! Hence the leet for "none" which is now "n0ne".) LQ 04:45, 30 June 2006 (PDT)

Actually, thinking about it, I don't think we've gotten any spamlinking since you turned it on. Probably too soon to really tell though. It's the 18-digit numeric strings that we're getting more of lately. Very annoying! So I added the $wgSpamRegex which currently blocks the displa of the css display code and a few phrases for some of the more popular commercial spam sites.

It relies on regular expressions, so, if we think it's safe, we could ban the regular expression for 18-digit numeric strings. Obviously it would have to be only applied to purely numeric strings. I don't want to do it if we can think of instances where legitimate 18-digit numeric strings might exist, but I can't think of any! Anyone else? LQ 04:58, 30 June 2006 (PDT)

Ah -- so I just put in a \d{20} into the regex (thanks Eric!) and I think it's working.

Now we just need to figure out how to edit the "Spam blocked" page which says your page has probably been blocked for external links ... it should say "or because you are presumptively spamming by putting in a meaningless string of digits"!!!!

It would be even better if we could block an edit if that were the *only* change. Right now, it blocks an edit if it *includes* a 16-digit string of numbers. LQ 11:20, 30 June 2006 (PDT)

External Links & References

• http://spamhuntress.com/2006/06/24/random-numbers-edits-on-wikis/
• http://meta.wikimedia.org/wiki/Talk:Anti-spam_Features