Trust management (TM) is an approach to access control in decentralized distributed systems with access control decisions based on statements made by multiple principals. Li et al. developed the RT family of Role-Based Trust-management languages, which combine the strengths of Role-Based Access Control and TM systems. We present a distributed credential chain discovery algorithm for RT 1 C , a language in the RT family that has parameterized roles and constraints. Our algorithm is a combination of the logic-programming style top-down query evaluation with tabling and a goal-directed version of the deductive database style bottom-up evaluation. Our algorithm uses hints provided through the storage types to determine whether to use a top-down or bottom-up strategy for a particular part of the proof; this enables the algorithm to touch only those credentials that are related to the query, which are likely to be a small fraction of all the credentials in the system.