Cleanup pages for pages that use 'term' without a language
It's pretty simple to construct using regular expressions. I used =German==([^-]|-[^-])+\{\{term(\|(?!lang=)[^\|\}]+)+\}\}
to construct one for German just using AutoWikiBrowser. (Did you want "term" in the title instead of "temp" or were you planning on expanding this to other templates that should have language parameters). It's probably better for someone to just run this regexp for languages for which there's cleanup interest (non-English would be the most bite-sized). Once the backlog has been whittled down then it can be easily be turned into a periodic, maintenance cleanup-list.
Oh yes that was a mistake, sorry. I meant 'term'. I can't use AWB because it's only for Windows, but I could probably do the same with Python. I don't really know how the XML dumps work though, could you explain it a bit maybe?
mw:Manual:Pywikipediabot has a module xmlreader that you can use to parse the xml dump into pages. Other tools are mentioned at m:Data dumps.