Abstract-Securely managing shared information in distributed environments across multiple organisations is a challenge. Distributed information management systems must be able to support individual organisations' information policies whilst securing global consistency and completeness. This paper proposes a multi-agent approach to a distributed multi-organisational system design based on this principle, focusing on the example of the distributed digital criminal dossier used in the Courts of Amsterdam and Rotterdam, compiled and managed by the Public Prosecution. Security requirements are identified and a distributed multiagent architecture proposed.