Résumé : Concurrent constraint programming (ccp) is a mature formalism for reasoning about concurrent systems that exhibit a constrained behavior. Spatial ccp and epistemic ccp are two novel variants of ccp currently being developed by Knight and Valencia. These variants model systems with spatial hierarchies of group information and knowledge. These systems are ubiquitous due to the advent of social networks and cloud computing where agents may share certain information with certain groups. This paper introduces an interpreter for these extensions we call k-stores. The interpreter is a Prolog implementation of the operational semantics of the languages allowing the programmer to simulate distributed information systems. The main feature consists of an implementation of a spatial (distributed) store that allows epistemic information in it. The system supports the specification of (named) processes along with the ccp classic primitives, namely, ask and tell operations. The declarative view of processes is inherited from the ccp extensions. The orthogonal implementation of the local space abstraction and the epistemic constraint system makes further extensions possible. Special attention is paid to the representation of distributed knowledge and common knowledge