Stakeholders' expectations of what an AI system can and cannot do should be managed so that trust is calibrated to actual capability.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.