There are lots of governments (including mine) that deliberately fabricate data, statistics, investigation reports to advance their agenda. How should we examine these claims?
If you have proof that a report is fabricated, with good references and all the good thingies we at skeptics like, feel free to challenge the claim that is presented on the report.
If you don't have proof - and by proof I mean more than "my government usually lies about stuff" - then you can't really challenge it. It would be your world against theirs, and normally a government has more leverage with skeptics than a single person.
This of course doesn't apply to philosophical or religious claims made by any government. If, let's say, a random christian nation says that 78% of teenagers would end up in hell because of some sort of gaming habit you can't really take the paper seriously for skeptics matters.
Otherwise, for police reports, economical data, census, etc. we can give a bit more of trust - not blind trust, mind you, but they are definitely worth considering and more powerful than some random blogger out there or some holocaust-denial nutcase screaming at the top of a soapbox.
As anything on skeptics - give us proof and/or references. If you have proof that a report is wrong, it will be welcome to the site. Otherwise, the community won't really take you seriously on your claims.
In questions, feel free to ask about those claims. Governments are generally not unique sources except some exceptional things.
I don't see a problem using government reports in answers either, unless there's reason to be skeptical about them (use voting to agree and disagree with the appropriateness).
I don't see how this is any different than any other topic on this site. You see a stat, are skeptical, ask a question. But you may have to give a decent reason why you're skeptical about official reports. I don't think "my government usually lies" counts. "Many have little faith in my government's reports, so third party verification would be nice" probably does count.