UTF-7 will certainly not be the only way to cause different interpretation by MediaWiki's parser and browsers. http://www.whatwg.org/specs/web-apps/current-work/multipage/semantics.html#charset has a nice overview of some other problematic charsets. Blacklisting them would certainly be a good start, however given the complexity of XML, I am not sure if it will reliably solve the issue.
Since Wikipedia serves its images from a separate origin, it is probably not seriously affected. For smaller wikis that do not have this separation, this results in an XSS attack.
Please attribute this to "Jan Schejbal / Hatforce.com" in the release announcement.