Another strawman. There is absolutely no requirement for such a constraint.
To demonstrate the validity of the biological ID hypothesis, it is enough to point out that specific examples of intelligent design -- in the form of large scale biological changes --can and do take place, on a daily and industrial-scale basis.
Validation of a hypothesis is not verification, of course (you do understand the difference, I presume). One cannot simply brush off the scientific necessity of producing evidence through testing.
Nevertheless, your suggestion that the post-production detection of such biological efforts is scientifically impossible, and therefore not worth doing, is rather difficult to take seriously.