A extension to the PrePost algorithm described by Yates et al. (http://core.kmi.open.ac.uk/download/pdf/5224046.pdf). The algorithm discovers static preconditions by querying Google for documents and creates feature vectors in order to learn a classifier for the static preconditions. The implementation is partial and to be used in conjunction with the LOCM algorithm, which generates domain models from partially observed action sequences, much like this implementation is intended to do.