We present Preambl, an approach that applies automatically generated invariants to regression testing and bug localization. Our invariant generation methodology is Precis, an automatic and scalable engine that uses program predicates to guide clustering of dynamically obtained path information. In this paper, we apply it for regression testing and for capturing program predicates information to guide statistical analysis based bug localization. We present a technique to localize bugs in paths of variable lengths. We are able to map the localized post-deployment bugs on a path to pre-release invariants generated along that path. Our experimental results demonstrate the efficacy of the use of PRECIS for regression testing, as well as the ability of Preambl to zone in on relevant segments of program paths.