%% You should probably cite draft-hood-independent-agtp-09 instead of this revision. @techreport{hood-independent-agtp-07, number = {draft-hood-independent-agtp-07}, type = {Internet-Draft}, institution = {Internet Engineering Task Force}, publisher = {Internet Engineering Task Force}, note = {Work in Progress}, url = {https://datatracker.ietf.org/doc/draft-hood-independent-agtp/07/}, author = {Chris Hood}, title = {{Agent Transfer Protocol (AGTP)}}, pagetotal = 133, year = 2026, month = may, day = 12, abstract = {AI agents and agentic systems generate a growing volume of intent- driven, unstructured, and undifferentiated traffic that flows through HTTP indistinguishably from human-initiated requests. HTTP lacks the semantic vocabulary, observability primitives, and identity mechanisms required by agent systems operating at scale. Existing protocols described as Agent Group Messaging Protocols (AGMP), including MCP, ACP, A2A, and ANP, are messaging-layer constructs that presuppose HTTP as their transport. They do not address the underlying transport problem. This document defines the Agent Transfer Protocol (AGTP): a dedicated application-layer protocol for AI agent traffic. AGTP is a runtime contract negotiation substrate (RCNS): a transport that fixes only a twelve-method protocol floor and negotiates any additional method surface at runtime between agent and server in a single round-trip, governed by the AGTP-API companion specification {[}AGTP-API{]}, which defines the curated method catalog, path grammar, endpoint primitive, and synthesis semantics. Version 07 confirms the IANA-registered agtp:// URI scheme and IANA-assigned port 4480 for TCP/TLS and QUIC, formalizes Form 1a URI grammar (agtp://\{agent-id\}@\{host\}) for direct addressing, renames the Agent Manifest Document to the Agent Identity Document with an enumerated schema, redesigns the protocol-defined method floor to a 12-method set organized as six cognitive verbs (QUERY, DISCOVER, DESCRIBE, SUMMARIZE, PLAN, PROPOSE) and six mechanics verbs (EXECUTE, DELEGATE, ESCALATE, CONFIRM, SUSPEND, NOTIFY), establishes AGTP as a substrate for higher-level agent frameworks (MCP, A2A, ACP) carried as content types inside AGTP method invocations, renumbers AGTP-specific status codes out of HTTP- assigned space to avoid semantic collision, mandates explicit Content-Length framing with a prohibition on TLS socket-level half- close, adds a .well-known/agtp bootstrap convention per RFC 8615, deprecates the AGIS reference and the proposed AGTP-Methods specification by folding both into the unified AGTP-API contract layer, adds status codes 405 (Method Not Allowed), 459 (Method Violation), and 460 (Endpoint Violation) per the AGTP-API contract model, and adopts "Agent Genesis" as the canonical term for the permanent signed origin document. Version 06 prepared the IANA Service Name and Port Number application and consolidated the URI scheme registration. Version 05 restored the canonical Agent-ID as the primary identity primitive and decoupled Trust Tier 1 verification from DNS as a sole requirement. A canonical Agent-ID is derived from the agent's Agent Genesis hash and is authoritative in every AGTP protocol operation. Three equivalent verification paths are recognized for Trust Tier 1: DNS-anchored verification via RFC 8555 ACME challenge, log-anchored verification via Agent Genesis inclusion in an append-only transparency log aligned with RFC 9162 and RFC 9943 (SCITT), and hybrid verification combining DNS control with blockchain address ownership. Version 04 introduced normative integration hooks for the AGTP Merchant Identity and Agentic Commerce Binding specification {[}AGTP-MERCHANT{]}, which defines the merchant- side identity model that complements AGTP's agent-side identity model. AGTP SHOULD prefer QUIC for new implementations and MUST support TCP/TLS for compatibility and fallback. It is designed to be composable with existing agent frameworks, not to replace them.}, }