A lot of research attention has recently been dedicated to multi-agent systems, such as autonomous robots that demonstrate proactive and dynamic problem-solving behavior. Over the recent decades, there has been enormous development in various agent technologies, which enabled efficient provisioning of useful and convenient services across a multitude of fields. In many of these services, it is required that information security is guaranteed reliably. Unless there are certain guarantees, such services might observe significant deployment issues. In this paper, a novel trust management framework for multi-agent systems is developed that focuses on access control and node reputation management. It is further analyzed by utilizing a compromised device attack, which proves its suitability for practical utilization.