Peter Ouyang just argued, during a phenomenology seminar, that pentaquarks do not exist - or, at least, they are not generic predictions of chiral soliton models. See experimental Fermilab news with the same conclusion. Peter's theoretical arguments were somewhat technical in character (well, some unconvincing pictures from the experiments definitely helped his case); some previous solutions were argued to disappear as soon as very small masses are introduced to the model and the SU(3) flavor symmetry is broken a little bit.
Pentaquarks in general
One of the interesting features of QCD is that all stable particles seem to be either 3-quark baryons, or quark-antiquark mesons. There are other ways how color-neutral states could be created, but these exotic hadrons usually seem to be absent from the spectrum of stable particles. This is why the object that decayed into a nucleon and the K-meson that contains an anti-s-quark was so interesting. Because of the anti-s-quark, there should be an anti-s-quark in the initial state. The processes are fast enough so that they must be caused by strong interactions which preserve strangeness; therefore, there had to be an anti-s-quark in the initial state, too. You may conclude that the minimal configuration in the initial state contained four quarks and one antiquark (namely anti-s-quark).
The question is whether QCD predicts such bound states. There are different approximate approaches to attack this question: SU(3) rigid rotator approach, quark models in general, the Skyrme model in particular, and so forth, and people talk about their different predictions. Neither of them is accurate and my feeling is that the qualitative physics question about the existence of the pentaquark is not really well-defined, at least not independently of the approximation scheme. Nevertheless, it seems obvious that the controversy will continue. And people have fun with it which is good.