[bawolff]

There is nothing to do, the incident was not caused by a vulnerability in mediawiki.

Basically someone who had permissions to alter site js, accidentally added malicious js. The main solution is to be very careful about giving user accounts permission to edit js.

[There are of course other hardening things that maybe should be done based on lessons learned]

[dboreham]

There are already tools and techniques to validate served JS is as-intended, and these techniques could be beefed up by adding browser checks. I've been surprised these haven't been widely adopted given the spate of recent JS-poisoning attacks.

[bawolff]

You mean like SRI? That's not really what happened here, so its not really relavent.

[streetfighter64]

Well, admins (or anybody other than the developers / deployment pipeline) having permissions to alter the JS sounds like a significant vulnerability. Maybe it wasn't in the early 2000s, but unencrypted HTTP was also normal then.

[bawolff]

That's a fair point, but keep in mind normal admin is not sufficient. For local users (the account in question wasn't local) you need to be an "interface admin", of which there are only 15 on english wikipedia.

The account in question had "staff" rights which gave him basically all rights on all wikis.

[cesarb]

> For local users (the account in question wasn't local) you need to be an "interface admin", of which there are only 15 on english wikipedia.

It used to be all "admin" accounts, of which there were many more. Restricting it to "interface admin" only is a fairly recent change.

[bawolff]

> Restricting it to "interface admin" only is a fairly recent change.

Its been 8 years!

[LaGrange]

> Well, admins (or anybody other than the developers / deployment pipeline) having permissions to alter the JS sounds like a significant vulnerability.

It's a common feature of CMS'es and "tag management systems." Its presence is a massive PITA to developers even _besides_ the security, but PMs _love them_, in my experience.