Agentic framework 2.1.2, agent identity: each agent should have a unique, cryptographically verifiable identity tied to a supervising agent, human user or department, with the capacity it acts in recorded, issued and tracked centrally to prevent agent sprawl; authorisations should be scoped, least-privilege, time- or session-bound and non-transferable with explicit escalation paths, and an agent acting for a human should never receive more permissions than that human holds, with delegations recorded.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.