In this paper, we propose an attribute-based access control model called ABAC–TG for online social networks (OSNs). This model comprehensively considers user and object attributes and two main social attributes: trust and gossip, which are calculated based on the Ego-node (the user sharing the information) point of view. Each user is evaluated trust and gossip wise by several criteria, such as total number of friends, number of interactions between two users, and more. A new algorithm for calculating user gossiping value by graph clustering is defined, and this gossiping value can also be used for trust calculation. The ABAC model is formally presented, including rules and attribute definitions, and is demonstrated by several use case scenarios. The gossip and trust assessments provide more accurate and viable information-sharing decisions that serve the purpose of more precise and flexible authorizations.This work is novel in two respects. First, we are using trust and gossip as dynamic attribute calculations. And second, we present a new algorithm for calculating the user’s gossip value from the ego user point of view and use it either as part of the trust attribute calculation or as a separate attribute in the ABAC model.