Here's an analogy to help better see why (B) doesn't work, why it's attractive, and what the difference is between a PSA/SA answer and an NA answer.
Premise: Tom is a cat.
Conclusion: Tom likes milk.
If you see something like the above in a PSA/SA question, you might anticipate an answer like (1) "All cats like milk." That certainly would help make the argument valid. But you also would not be surprised to see an answer like (2) "All mammals like milk." Since that too would also make the argument valid (under the reasonable common sense assumption that all cats are mammals). In other words, both (1) and (2) could be the correct answer choice for PSA/SA questions.
However, just because (2) helps the argument does not mean that the author of the argument assumed it. The author could easily say, "No, I wasn't thinking about mammals at all. I was only talking about Tom, cats, and milk." It would be unreasonable to claim that the author assumed anything about mammals even though assumption (2) helps the argument greatly. Such is the nature of very strongly helpful assumptions.
I suspect this confusion might be what tempted many of you to choose (B).
Analogously, if you restate (B) to say "anyone whose political motivations are clearly discernible is an unreliable source of information to legislators", you'd get a correct PSA answer. (B) shoved back up into the shitty argument in the stimulus would really help the argument out just like how (2) shoved back up in to the Tom/cat/milk argument would help that argument out. But you cannot say that the argument assumed it. That's the difference. (B) is not descriptively accurate.