What happened
The individual Internet-Draft 'Agentic AI Use Cases and Requirements' (authors Reddy.K/Sarker/Reddy of Nokia, Yao of China Mobile) was updated to revision -03 on 9-10 October 2026 (datatracker: last updated 2026-10-09; document dated 10 October 2026, expires 13 April 2027). Verified by fetching the datatracker page. The draft describes use cases for agentic AI communication systems (single-agent tasks, orchestrator/agent collaboration, long-running delegated tasks with authorisation checkpoints, peer collaborative multi-agent problem solving, cooperative reasoning/consensus, tool/data/API mediation) and derives protocol requirements intended to guide IETF standardisation on agent-to-agent and agent-to-tool communication — dialog management, transport, multimodal communication, communication security, agent identity and authentication.
Why it matters
This is a fresh, in-window IETF output defining protocol-level security requirements (communication security, agent identity and authentication, authorisation checkpoints) for inter-agent and agent-tool traffic. It sits alongside the already-tracked agent identity/permission-receipts/audit-trail drafts and signals where future standards for secure agent communication are heading, relevant to anyone building interoperable agent protocols.
Action needed
Track the draft for WG adoption; review the derived protocol security requirements (identity, authentication, communication security) when designing agent communication and authorization-checkpoint controls.