MediaWiki Administrator's Handbook/Combating spam

From Wikibooks, open books for an open world
Jump to navigation Jump to search
Pages to watch MediaWiki Administrator's Handbook
Combating spam

Spam on Wikimedia projects[edit | edit source]

Wiki spam is often thought of in it's more blatant form. Unwanted links with keywords such as eBay! or Cialis!, but that blatant form is actually comparatively rare on wikimedia projects. If you're interested in learning about combating spam on your own MediaWiki installation (which often takes this more blatant form), refer to the earlier section: MediaWiki Administrator's Handbook/Spam and Spammers

On Wikimedia projects "spam" is instead defined as the behaviour or pushing links inappropriately, and this can take far more subtle forms. For example it is entirely possible that the domain could be spammed (though the domain itself is not spammy). Domain owners will often be outraged (understandably so) that their reputation has been damaged by someone else spamming the domain.

Adding external links to an article or user page for promotional purposes is disruptive, and is considered to be spam. Although the specific links may be allowed under some circumstances, repeatedly adding links will in most cases result in all of them being removed. It is the behaviour in placing the links rather than the content of the site that may cause concerns.

As Wikimedia projects use nofollow-tags adding a link to one of them doesn't alter the search engine ranking.

Detecting spam[edit | edit source]

Spam can take one of the many forms possible, like

  1. A username that clearly looks promotional or refers to that of a company and its edits are done just for that (often this happens on the user's main page)
  2. Text in a language that is not the main language used by the project
  3. Inserting gibberish into many pages
  4. Creating multiple pages with the same or relatable content.

One-wiki spam[edit | edit source]

The first step is to delete all the page(s) which were created by the spammer. If there are a lot of pages, use Special:Nuke to delete it all at once. If it's all inside existing pages, revert that user's edits manually.

The next step is to block the user. Usually that block will be indefinite. You may revoke email access, but be careful of revoking talk page access as it prevents users in good faith who may have mistakenly spammed (though do it if there's evidence of talk page spamming). Keep autoblock on; spammers generally try other ways to continue spamming. If multiple accounts are spamming, set the block to affect logged-on users as well.

Cross-wiki spam[edit | edit source]

Cross-wiki spam means that a certain link has been added to multiple projects, normally by a single user, a set of SPAs, or an IP range. In this case the URL might be blocked (blacklisted) on a global level, meaning that it can't be added to any Wikimedia project. Global IP blocks may be placed to augment the blacklisting, for example, when malicious spammers rapidly switch domains, or abuse lag time on the blacklist.

Combating spam[edit | edit source]

Method 1: Revert and remove the spam[edit | edit source]

Method 2: Block the user(s)[edit | edit source]

When one or a few users or IPs are adding spam links, they can simply be blocked.

For cross-wiki spam, global accounts may be locked where possible, and short global IP blocks can be considered in severe cases. Blacklisting is the main method of combating cross-wiki spam.

Method 3: Page protection[edit | edit source]

When one or several pages are targeted by a spammer, they can be semi-protected to stop anonymous and new users from editing, or full-protected if established accounts are involved. However, blocking is preferable unless the spammer is making new accounts to spam the wiki, since page protection affects legitimate users as well as the spammer. If a clearly bad page (like obvious spam titles) is being repeatedly created, create-protecting the page could be a better idea, but do not do this if only the content is spam.

Method 4: Blacklist the domain(s)[edit | edit source]

Local blacklist[edit | edit source]

Each Wikimedia project has a local blacklist and a local whitelist, for example en:MediaWiki:Spam-blacklist and en:MediaWiki:Spam-whitelist. The former can be used to disallow linking to certain domains on a certain project only. The latter is a means to be able to link to domains on the global blacklist, or to exclude certain documents on a server, while all other documents on that server can not be linked to.

Addition and removal of links can most often be discussed on the local talk page. Before blacklisting all other measured should be explored such as blocking the user or IP or protecting the page. However where this is not sufficient local blacklists should be used whenever possible to manage blocking of external links; the global blacklist should only be used for widespread cross-wiki link placement, where other local options are not effective. The list may also be used for sites that simply are not suitable for whatever reason if there is consensus.

Global blacklist[edit | edit source]

There is a global blacklist on Meta. This is a list of domains that have been placed excessively, or which are deemed completely unhelpful to any project, and are therefore blocked on all Wikimedia projects, i.e. you can't link to one of the listed entries from any Wikimedia project. For a domain to be globally blacklisted it generally has to be excessively linked on multiple projects. If it's only added to a single project generally the link should be blacklisted locally (some exceptions exist).

Addition and removal of links can be discussed on m:Talk:Spam blacklist.

Exceptions[edit | edit source]

Some links are generally blacklisted on meta, even if the abuse has only been to one project, or when the link has not been used abusively yet:

  1. URL-shorteners/redirect sites (like e.g. tinyurl) as these can be used to circumvent blacklisting of other domains, and it is totally unnecessary to use these (as one can link to the original document directly).
  2. Sites which drop viruses, Trojan horses, etc. etc. on the users computer, in other words, links which, when clicked on, are a threat to the computer the user is on. Blacklisting of these sites is generally temporarily, until the problem has been resolved.

Helping out[edit | edit source]

If you're interested in helping out by fighting and removing unwanted external links, please see m:Spam blacklist/help. To get started, you may wish to join #wikimedia-external-links or contact one of the following users: