Hiding Information in Multi Level Security Systems

In this paper we analyze the possibility for malicious agents to transmit an information possibly hidden in a Multi Level Security System via a covert channel. We give a framework for which we get two decidability results. Firstly, given a code and a system one can decide whether the system allows a covert channel for this code. Secondly, one can decide whether there exists a code to transmit one bit of information, the code is computable as well as the strategies of the two partners.