An Argument of Circles
If you are a naturalist, believing that nature is all that exist, please do not try to argue to that point by filtering all evidence through methodological naturalism. Methodological naturalism demands you only allow natural causes into your explanations. As such, this argument form is invalid. Specifically, it is a circular argument also known as Petitio Principii. This Latin is sometimes, wrongly, translated as “begging the question.” But a better translation is “assuming the initial point.” In other words, one is assuming the conclusion to prove the conclusion. In this case, the naturalistic argument is circular because of the role of methodological naturalism in the argument. If the method of argument presupposes the conclusion then it’s not really an argument but, instead, a fancy way of repeating oneself.

It’s not much of an argument to say, “The Bible is true because it says it’s true, therefore it’s true.” That’s really just a repetition of the claim, “The Bible is true.” Likewise, it’s an equally circular argument to say, “There are no supernatural causes because when we assume there are no supernatural causes we find there are no supernatural causes.” Welcome to the argument of circles.

Think of it this way. Suppose you are a prosecuting attorney and you are trying to argue that the defendant is guilty of murder. You may have any number of evidences to offer in the case, but as it turns out the judge tends to think that the defendant is innocent because the defendant has been innocent of most murders in the world and everyone else alive has yet to be murdered by this guy. The likelihood of his guilt is roughly 7 billion to 1. Moreover, defendants are considered innocent until proven guilty. Therefore, the Judge concludes, that he is most likely innocent of this murder in question. The judge, then, adopts a method of trial which will hear only the defendant and his evidence (including witnesses). Meanwhile, you and all of your evidence (including witnesses) are not even allowed in the courtroom. You are not allowed to interrogate witnesses, introduce contrary accounts, question material evidence, introduce evidence of your own. And even if you could sneak some of those evidences into the courtroom, the judge demands that all accounts of that evidence adhere strictly to the assumption that the defendant is innocent. You are not even allowed to sit in the same courtroom with these guys. The entire contrary view that you represent–the attorney, witnesses, material evidence–is methodologically excluded from the trial.

Is that a fair trial? Of course it’s not a fair trial. One cannot render a sound verdict if only one side is allowed to speak. In this way, methodological naturalism is not a suitable all-encompassing method to assess whether supernaturalism is true. To carry this point a bit further, methodological naturalism is not even a good way to assess if naturalism is true either. It never entertained any probability for anything other than (metaphysical) naturalism. In such cases, one cannot draw responsible and safe conclusions about the probability of something if one’s method presupposes a 100% likelihood. How likely is naturalism if the very question requires that we assume naturalism must be true? Well, it’s 100% likely, not because of evidence or argument or of how reality is but because one has stipulated that to be case by force of method. The trial is rigged to achieve one conclusion.

Perhaps Methodological Naturalism’s Usefulness Redeems It Somehow.
Methodological Naturalism will work pretty well if nature is all that exists, but honestly, a great many supernaturalists affirm an orderly world where natural causes operate reliably and knowably and can safely affirm the general usefulness of methodological naturalism. The general utility of methodological naturalism is not big threat to supernaturalism since, at least among the Abrahamic faiths, supernaturalists can affirm all the natural causes, an orderly world, knowable processes to the universe, etc. and still believe in a more ultimate creator and sustainer God somehow overseeing all of it.

Perhaps Methodological Naturalism Can Be Argued On Other Grounds and Then Used to Support Naturalism
The best option remaining seems to be arguing first for methodological naturalism and then using that method and its evidences to argue against supernaturalism. Here you can muster a logically consistent and evidentially coherent case against supernaturalism. In this way, the overall argument that includes methodological naturalism has not filtered everything through Methodological Naturalism. Possible supernatural causes were allowed into the courtroom in the pre-trial, so that some contrasting views or counter-evidence were offered there, and a method of trial was adopted there. For example, the murder suspect turned up as a witness saying, ‘I’m not dead.” If that happened in the pre-trial phase, then the rest of the trial–one would imagine–would go very differently. that would not be a circular argument, but neither would it be using methodological naturalism to frame the everything else.

Conclusion
To summarize things so far, there is a valid way to employ methodological naturalism within an overall argument for (metaphysical) naturalism. One must argue to methodological naturalism, one cannot simply assume it, and then use it to prove that there are no other causes. That method prohibits you from allowing any others into view. Naturalism can be argued. But to keep such arguments from devolving into a circle one cannot safely assume methodological naturalism throughout the argument. If everything is filtered through methodological naturalism then all possible supernatural causes are presumed false without trial, without consideration, without being allowed even into the courtroom. To render a verdict on that basis is a miscarriage of justice.