The paper addresses problems of efficient knowledge representation and exploration of social knowledge enabling agents' integration into the multi-agent community.The proposed solutions are based on task decomposition, organization of negotiations, responsibility delegation and other aspects of social behavior in multi-agents systems aimed at integration of already existing software components.It is supposed that all the social knowledge is kept separated from the problem solving knowledge and agents' specific internal intelligence and that it is organized and administered in the acquaintance models located in the agents' wrappers. A specific tri-base acquaintance model is formalized and discussed throughout the paper.This model helps to optimize the communication traffic and to implement meta-reasoning processes.Practical applications of the acquaintance model in different fields are presented, acquired experience is discussed.