As for the rule resolving possible conflicts in aspect between a verb and its object sentence, perhaps that’s the reason, but I don’t see why the aspect of a verb taking an SAO and the verb in its SAO have to agree in the first place. If I say (ungrammatically) {wa'leS SIStaH 'e' vI'aqpu'} “I have predicted that it will be continuously raining tomorrow”, in both Klingon and English the prediction is completed and the raining is continuous, and there doesn’t seem to be any reason grounded in laws of nature why the aspect of these two verbs must agree. Even with {qama'pu' (vI)jonta' (vI)neH}, there’s no reason the wanting has to be completed. {Qugh HoD} could still not-completed-want to capture prisoners, even if the opportunity to do so is apparently no longer available.

If I have to just say {wa'leS SIStaH 'e' vI'aq} or {wa'leS SIS 'e' vI'aq} instead, since verbs taking SAO can’t also have a VS7, then the information that {'aq} is completed has been lost, and if it can be lost here, why not somewhere else? Certainly, the aspect of the verb taking the SAO doesn’t come from the verb in the SAO, since {wa'leS SISpu' 'e' vI'aq} sounds like the prediction is that it will stop raining tomorrow, when what I really want to say is that I finished predicting that it will rain tomorrow. (I’m now fairly convinced that you do indeed need to use aspect markers when the meaning calls for them, so I’m not holding this up as an argument to say they you don’t; I’m just trying to understand the ramifications of this restriction more fully.)

As somebody who is familiar, to varying levels of proficiency, with languages that indicate aspect but not tense, tense but not aspect, and both tense and aspect, I’m aware that dropping aspect markers when they would otherwise be indicated, in languages that mark aspect, produces speech that seems as off as “I go to the store” (when I in fact went to the store) does in English, but just because that’s true for other languages that mark aspect that I’m familiar with, I didn’t take that to mean that it’s necessarily true for Klingon as well. I do make an effort not to allow my intuitions from other languages creep into my understanding of new languages I learn, although to a certain extent it’s unavoidable, I suppose. It’s hard to say for sure, but I think my personal impression of the aspect markers being mostly optional was mainly informed by being exposed to usage of non-aspect-marked verbs to indicate what appeared to be completed actions (problably mostly from Duolingo) early on when studying the language, which fed into and reinforced my (erroneous, as SuStel points out) view of Klingon as communicating “no more, no less” than necessary.