It is at least logically possible that the universe be designed in such a way that the designer is undetectable. It's an internally consistent hypothesis, and it's consistent with everything we observe. However, it doesn't follow from that that it's a reasonable thing to believe.
Other more radical hypotheses, like occasionalism are be internally consistent, too. The problems with admitting such beliefs to our storehouse of beliefs often lie not in that they contradict themselves or that they are entirely inconsistent with what we observe. Rather the problems often lie either:
in reconciling such hypotheses with our other beliefs. For example, we may believe that mutation is genuinely random in relation to natural selection, and if so, it's hard to square that with genuine design. Or,
in trying to determine how to cash out the hypothesis in terms of particulars, and then finding any justification for that whole package. Once you start trying to think about who or what this designer is, you typically find yourself having no grounds for justifying any particular way of cashing it out. If the existence of a hidden designer is consistent with the evidence, it doesn't follow that the claim that a designer exists — or some particular kind of designer exists — is especially plausible or justified. For instance, it might be equally well justified that we are all a feature of the imagination of Descartes's evil demon. In other words, even if you can make some view (designer, or evil demon) consistent with what we observe and understand, that's a very long way from justifying it as a reasonable thing to believe. Of course I don't believe that we're a figment of the imagination of an evil demon, and I don't think it's reasonable to believe that; and yet it's completely consistent with the evidence. And that consistency suggests that epistemologically we shouldn't be over-impressed by a hypothesis not contradicting our other beliefs. I do not feel forced the consistency of my beliefs with the evil demon hypothesis, that is, into agnosticism about evil demons. To reach that point, I would need some significant further epistemic motivation. And in the case of design, that's been lacking.
That is to say: while there can be no definitive counter-evidence, and while there are a number of ways the hypothesis can be fit with our knowledge, it can only be done quite awkwardly. We should take that awkwardness of fit as a strike against the hypothesis. This is how science works more generally: we don't arrive at complete certainties so much as deeply-corroborated sets of coherent beliefs.
But also, we should not understand lack of evidence against something as evidence for it. I look outside, and I'm alarmed that I have no evidence that there's not a hobbit outside my window wearing an invisibility ring. I literally have no evidence that he is not there. But this is not positive motivation for believing that he is there. (If it were, it would also be positive motivation for believing that infinitely many other invisible things are outside my building: invisible orcs, invisible dragons, invisible ironing boards ...)
Then, without some positive motivation for believing something, there is no reason to be agnostic about it. "Agnostic" is explicitly defined in various ways, but I take it you mean that the existence of a designer remains an open question. I would suggest that without positive motivation to believe something, we shouldn't treat them as open questions. They're not closed questions, either, but there we have literally no reason to continue to ask the question. This applies equally to the idea that a designer designed the universe then covered its tracks and to there being a dozen dancing invisible ironing boards outside my window. No evidence suggests they couldn't be there, but by the same token nothing raises the question of whether they are there. Nothing makes it a question worth spending time on any more than whether there are half a dozen invisible ears of corn.