Attack Models. Potential attackers and their motivations are identified so threats can be modelled realistically.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.