Skeptics get a bad rap.
We need to differentiate between a Skeptic worldview (where we flat out reject everything) and a Skeptic epistemology, where we demand verification of any and all truth claims (both religious and otherwise). Skeptic epistemology is a moral imperative (religions- particularly Christianity- agree with me here. In general, they contain lots of warnings about false prophets and changes to scripture). I want to talk about what I see as the proper application of skepticism in our search for truth.
The first thing we have to realize (or rather admit) is that our perception of reality is untrustworthy. There are too many studies on human psychology that show us doing horribly irrational things for us to entertain the notion that what we think we know always matches with reality. Moreover, there are too many painfully real present-day examples of people doing absurd things in the name of their religion. So when is it appropriate to trust our senses, feelings, and beliefs, and when is it not?
There are a few instances that jump to mind here. If your belief in something is based on feelings, emotions, Faith, Divine Impartation of knowledge, an Infallible Leader, or any other objectively unverifiable basis, you have a problem: it's a virtual certainty that someone, somewhere (probably a great many someones) believe in a diametrically opposed truth based on the same evidence. The Christian claims divinely imparted belief? So does the Muslim. The Buddhist claims inner peace and joy as evidence of his religion? So does the Christian. Your specific belief may be unique, but the basis of your belief is not. So on what strength are you to say that your belief is well founded, but everyone else's is purely psychological? It seems like you're forced to pick one of two options- either everyone else is lying, or humans are capable of being utterly convinced of a truth (on the same basis that you are utterly convinced), and still be wrong.
If a claim demands action but makes no prediction, you ought to be extremely suspicious. Such claims have historically been used as methods of control, but more to the point, you don't actually have any reason to believe them. Once they start making proscriptive claims about what you ought to do, they are now literally asking you to bet something on their truth. It's one thing to accept a claim as plausible (and "believe" it, in the weak Bayesian sense) on the strength of authority- you believe a friend who tells you your favorite sports team won. But once you have to take action on this belief (or abstain from action), your standard for belief (your critical Bayesian level, if you will) must increase in proportion to the level of action required. If a physicist tells me certain kinds of radiation are not harmful, I believe him. If he asks me to carry some of the glowing material, the standard of evidence I require increases a great deal.
The key thing i'm arguing for here is that we have to recognize the fundamental possibility that our truth-telling mechanisms are wrong. I say "fundamental" here in the sense that this possibility is always present, no matter what level of enlightenment, relationship with God, understanding of the universe, or epistemological epiphany we reach. Not only is it always possible that we're wrong, it's actually statistically likely that we're wrong. A claim of anything else is either a claim that you believe your truth more strongly than any other religion believes in its truth, or that you have a better truth-telling mechanism than they do (a claim that suffers greatly when your belief depends on Faith, emotion, etc). Science, it should be noted, claims the latter- that it has a better truth-telling mechanism (it's a pretty convincing claim).
The point here is that, as the relative extremity of the action required increases, so too should your requirement of evidence. And in fact, there are some things that might be true, but you should be so suspicious of them, that in practice you never actually believe them. If a religious claim ultimately requires you to commit what you know to be a moral evil, you are obligated to reject the religious claim- even if it's true! Because our truth-telling mechanisms are untrustworthy, it's much, much more likely that you're wrong about what God is telling you than it is that God is actually telling you to slaughter all those innocent people.
So what's the takeaway here? First, most people aren't skeptical enough. Second, Faith is not valid evidence. Third, whatever belief system you end up with, it must be beholden to your moral principles, not the other way around. It is wrong- both ethically and epistemologically- to subjugate your conviction of right and wrong to any authority other than your own. You are responsible for your actions. Should such a day come, you will stand before God on judgement day. And I can think of no other response that a benevolent God could have to someone who knowingly did wrong in his name than "How dare you?"