According to this paper, I am heading in the wrong direction: http://www.unc.edu/~maguilar/UNCNCSU/NIG_EGW.pdf
Binomial model with skewness and kurtosis code p 297 p 298. It may be worth you try with a trinomial model. The implementation is straightforward.
"For american pricing you just need to use Haug's code with skewness and kurtosis if you want with the american option feature, the boundary condition : an america call value is max((S-K); same european node value)) everywhere on the tree ( S=spot K= strike). The same for american put pricing." You just need to improve Haug's european option code by implementing the american option feature into the code. But if you understand what make the difference between european style and american style option in binomial model, hence you understand how to write the code for trinomial and finite difference schemes. If you derive a binomial tree with skew and kurtosis for european style option, you just have to add this constraint that at every nodes the option price is max(....).
Ah! I understand what you mean now. I thought that is what you were trying to say, but I wasn't sure. Thanks.