Heterotic orbifolds continue to provide an attractive framework for top-down, globally consistent string model building. Having identified explicit models that give rise to potentially realistic particle spectra (no chiral exotics), one would now like to be able to study dynamical issues like moduli stabilization and decoupling of exotics in these explicit models. This requires the low energy effective field theory
action for a given compactification. Towards these objectives, we have revisited the question of allowed superpotential couplings in the
low energy effective field theory via the string CFT, and found some new selection rules, as well some corrections to previously known
rules. I will review our computation and results.