Depoe on Internalism and the AFR
I don't have Hasker's book with me at the moment, but Hasker does say that a thoroughgoing externalism would be incompatible with his version of the AFR. Though on my multi-track model Plantinga's argument would be a way to get the argument going if you were an externalist.
Still, the argument is not about general theories of epistemic justification, but is rather about the fact when an atheist uses the argument from evil, or a scientist uses a mathematical equation to support a thesis in math, in order for it to be what the arguer from evil says it is, or in order for it to be the kind of mathematical inference the scientist says it is, there has to be an understanding of the propositions, a perception of a logical rule, and the reaching of the conclusion through a process of inference, the perception of what Lewis calls a ground-consequent relationship. Simply having a "black-box" reliable belief-producing mechanism is not enough to make these processes "as advertised." Internalists are good at these kinds of inferences. What they are not so good at are cases like, well, Steven Nash knowing when to pass the ball to Shawn Marion to get a dunk against the Spurs. :)
I'm not sure I have a general theory of epistemic justification that covers all cases. But reliabilism doesn't cut the mustard in the cases of scientific and philosophical inference. Link