The problem with this statement is that you're a priori assuming that we could understand neither the methods nor the motives of the "unspecified agent." As a result, you reach the untenable conclusion that it is impossible to bound our observations of the putative agent's handiwork.
In reality it is not "obvious" at all. The physical nature of the systems in question tends to bound the solution space to a large extent. Further, human experience with those systems provides a means of understanding the limits of the solution space, as well as a ready-made set of techniques that can potentially be tested for.
No, the reason why we exclude ID if nothing is known about the supposed designer is the fact that it is not falsifiable.
And thus neither is its alternative.
And thus neither is its alternative.
If by this you mean the TOE is not falsifiable, you are quite simply wrong. The TOE is readily falsifiable. However, no evidence has ever been presented to falsify its tenets in any major way.
It is falsifiable because it makes predictions both negative and positive. It makes predictions about what you cannot find if the TOE is accurate, and it makes predictions about what you should find if it is. So far, nothing has been found that clearly shouldn't be there, and everything that has been found has accorded pretty well with the positive predictions it does make (I'm not even going to bother regurgitating viral insertion studies and fossil strata predictions -- that's been done ad nauseum on other threads and is readily accessible).
Well, that automatically follows from the unspecified nature of the alleged designer. As soon as we have the designer, I don't deny that we can learn about his methods and motivations and make inferences on what we should resp. shouldn't observe.
However, as long as this designer remains unidentified we can't make any meaningful inferences.
And to be honest, I haven't seen that any IDer has in any way identified this alleged designer or ascribed any concrete properties to him.