Abstract Meaning Representation (AMR) is a representation for open-domain rich semantics, with potential use in fields like semantic parsing and machine translation. Node generation, typically done using a simple dictionary lookup, is currently an important limiting factor in AMR parsing. We propose a small set of actions that derive AMR sub-graphs by transformations on spans of text, which allows for more robust learning of this stage. Our set of construction actions generalize better than the previous approach, and can be learned with a simple classifier. We improve on the previous state-of-the-art result for AMR parsing, boosting end-to-end F1 from 59 to 62 on the LDC2013E117 and LDC2014T12 datasets.