If someone knows that Q, that person cannot be thrown into a state of epistemic confusion regarding Q.
I doubt this. It seems that in many cases knowledge can be undercut when we acquire further (perhaps misleading) beliefs. I may see (and hence know) that the widgets are red, until I form the defeating belief that they are irradiated by red light. Similarly, if an apparently infallible source ('God') tells me that I can't know that Q, his assertion may be self-fulfilling, by raising doubts and plunging me into a state of epistemic confusion and uncertainty that is incompatible with full-blown knowledge.
On the other hand, the unrestricted claim that epistemic doubts are self-fulfilling leads to contradiction when combined with other plausible assumptions (introspection and closure under known entailment). Perhaps the two most promising options are then:
(1) Allow only justified doubts to act as defeaters; or
(2) Restrict the claim, so that only undoubted doubts are automatic defeaters. If you come to believe that your doubts [about your belief that p] are unjustified, then those doubts no longer undercut the justification for your belief. (The defeater is itself defeated.)
Remaining questions: Which of these options is right? If (1), will that make some knowledge indefeasible after all? Or is the doubt and confusion raised by God's testimony justified, so that it may defeat your knowledge in any case?