The SCOOP model provides programmers with a simple extension of Eiffel that allows them to produce high-quality concurrent applications with little more effort than sequential ones. The model is simple yet powerful; nevertheless, its access control policy is pessimistic: (1) all separate actual arguments of a feature call are locked, even if it is not necessary, and (2) at most one client object can access a given supplier object at any time. As a result, SCOOP-based programs are deadlock-prone. Additionally, some interesting synchronisation scenarios, e.g. reader-writer, cannot be implemented efficiently. This paper presents an extended access control policy for SCOOP that makes it possible to specify which arguments of a routine call should be locked; it also allows for shared access to supplier objects thanks to the interleaving of pure query calls. Additionally, we introduce a mechanism for lock-passing that increases the expressive power of the model, thus allowing the programmer to write programs that were impossible to implement in the original SCOOP model.