{"version":"2.1.312","anchor":"agent-protocol-schema-gains-new-client-and-agent-entries","canonical_anchor":"agent-protocol-schema-gains-new-client-and-agent-entries","heading":"Groundwork for consent and resume handling between connected agents","tier":"internal","area":"Multi-Agent","scope":"individual","heads_up":false,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.312\/e\/agent-protocol-schema-gains-new-client-and-agent-entries","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.312","markdown":"### Groundwork for consent and resume handling between connected agents\n\nClaude Code's internal agent messaging format gained fields for refused messages, replies awaiting consent and resume position\n\n**Unclear.** It is not clear what feature these fields belong to or whether any of it can be reached by users yet.\n\n**What**\n\nClaude Code has an internal message format that lets it exchange messages with other agents and clients. That format gained one new client entry and one new agent entry, and several existing entries were replaced. New fields track the state of incoming messages:\n\n- whether an incoming message was refused\n\n- which connected peers were restored\n\n- which replies are waiting for permission before they are sent\n\n- where to pick up again when a session resumes\n\nThe saved resume state also gained a field that records a question asked by a subagent (a helper agent Claude Code starts to handle part of a task).\n\n**Why**\n\nThese fields suggest that work is under way on asking for consent before replying to messages from other agents, and on resuming those exchanges. Nothing in this release shows such a feature being visible to users yet.\n\n- Area: Multi-Agent\n- Tier: Under the hood\n- Useful: 3\/5\n- Signal: 4\/5\n- Scope: individual\n- Heads-up: no"}