In today's Online Social Networks (OSNs), one fundamental issue is to allow users the right to monitor the messages posted on their own private space and to avoid posting inappropriate information. Virtual Social Network so far offers no support for this criterion. Within this paper, to fill the void, we suggest a framework that enables users of Online Social Networking to have full control of the messages posted on their walls. This is accomplished through a versatile rule-based system that enables users to configure the filtering criteria to be applied to their walls, and a soft classifier based on Machine Learning that automatically performs message labeling to facilitate content-based filtering.