Mark,
You laid out the case for Agatha being a daughter of Stephen very nicely, but you missed the main point -- similarly impressive cases can be made for the other theories as well.
Earlier, you characterized the Hungarian filiation as being the "traditional" descent. Not so. The traditional descent is that Agatha was daughter of a brother of Henry II. The Hungarian descent is a modern replacement.
You also said that, "I do not see that the different genealogical theories about Agatha have really examined the original documents and their provenance." Also not true. Ronay, de Vajay, and Jette have all written in-depth analyses.
The case for Hungary is essentially this -- Edward was in exile in Hungary, so his wife was most likely Hungarian. Plus, one of the early chroniclers said she was a sister of the Queen of Hungary. And, I agree that's a very good point.
One of the arguments on the other side(s) is that if Agatha had been the daughter of any king, anywhere in Europe, it would have been stated from the first. We wouldn't have obscure references to her as a niece of Henry II, or a sister of the Hungarian queen. Those are, according to this argument, evidence that her father was someone who would have been unknown in England. So, her parents were skipped over in favoring of naming a famous relative.
One of the best candidates from that point of view is Liudolf, the margrave of West Friesland, who was a half-brother of Henry III. Liudolf is particularly attractive because he is known (somehow, I'd have to look it up) to have been a close relative of Agatha. Henry III's daughter Judith (perhaps Agatha's aunt) married Salomon of Hungary, so we're back to a Hungarian connection but this time without a royal father.
The problem there is that Agatha's name is Greek, and Agatha and Edward gave their two daughters names that were then Greek (Margaret and Christine). Those names appear at this period only in Sweden. If Agatha was Liudolf's daughter, why is her name Greek, and why did she give her daughters Greek names? So, there is an argument that Agatha must have come from a court influenced by Greek culture -- which really means Kiev. Yaroslav famously promoted Greek culture, and used his step-mother's Byzantine connection to enhance his own prestige. It was from the Kievan court that the Swedes of this period got their Greek names. On the other hand, Hungary was in the cultural orbit of Germany, as you can see from their names and marriages.
Since you're a medievalist too, you probably know that the "plain meaning" of a medieval Latin phrase is an unimpressive argument. Medieval Latin is debased compared with Classical Latin. The people who wrote Medieval Latin were often both unskilled in composition and trying to turn an elegant phrase. We have no shortage of genealogical problems caused by the vagaries of Medieval Latin.
The arguments go on and on, back and forth. Not everyone here would agree that we are doing academic genealogy, but if we were we'd have to accept the general opinion of modern scholars that, despite many interesting theories and arguments, Agatha's parentage is unknown.