Without taking sides between Krusty and Bodhi, the problem seems to stem from a simple misunderstanding of how "stealth" works (or in this case didn't). BTW, none of what I'm going to say is classified, it's basic Stealth 101 and is easily available in open source plus these fundamentals have been known for decades. There is a lot more I'm not saying but the fundamentals is all that's important to this discussion.
Stealth airplanes are not completely invisible to radar and how visible they are depends a lot on radar frequency. Only so much radar NRG can be absorbed, the rest is "managed" by reflecting it away from the emitter. The B-2 is an easy example. If you look at the top view of the plane you immediately notice that it's angular...I know...duhhhhh. Those angles serve to reflect radar NRG off at a different angle than where it was received from but there's a problem with this. Each of the leading and trailing edges (and even seams along panels and edges) creates a relatively strong reflection to any source that's perpendicular to it because each is basically nothing but half of a dipole antenna. This is particularly a problem if that source is low-frequency (i.e., long wave-length) with a wave length the same (or multiple of) the physical dimensions of the object it's hitting. Since most airplanes are designed around aerodynamic, not radar reflection, considerations there can be hundreds of straight edges and each creates its own reflection so the average airplane is one big old radar reflector and is easily detected from any direction. This is a problem even if the whole vehicle is covered with RAM so the designer has to manage the reflection that can't be absorbed. Notice in the picture of the B-2 how almost every straight edge or seam on the plane is parallel to the leading edges of the wings.
In other words every single edge is perpendicular to only four distinct directions so the B-2 has only four lobes perpendicular to the leading and trailing edges which means that those are points at which it's easier to detect. I'm not saying it's easy to detect, just easier compared to other directions. There's still lots of other things going on but this is basic physics. That's also why the F-117's construction is only part of how it achieves stealth, the other is a mission planning system that manages it's route so as to control where those reflections are pointed.
The fact that the F-117 in Serbia flew a repetitive track made it predictable. The fact that the Serbs moved their radars made them unpredictable. When you combine their apriori knowledge of where the F-117 would go along with some weak detections by their low-frequency, long wave length air search radars it's fairly easy to see how they improved the likelihood that they would get a shot. Also, based on the fact that the F-117 pilot didn't get any RAW indications means (to me) that the missile was optically guided which is made possible because the missile operators knew where and when to look for the target. The claim by the Serb commander that he modified his antennas seems extremely unlikely to be true. He made have tried some mod but the fact that he successfully killed the F-117 doesn't prove the mod did anything at all. It's sort of like those people that put one of those supposedly power enhancing things that plug into your car's cigarette lighter and then claiming it gave him 30 more HP. Bull. The only thing he could have done that might have been even remotely successful would be to tune his radar frequency to give him a wavelenth that was an exact multiple of the F-117's leading edge but I have no idea is this is even possible with his equipment.
