On July 23, at IETF 126 in Vienna, the Internet Engineering Task Force held its first Birds-of-a-Feather session dedicated to standardizing how AI agents talk to each other and to tools. The session, called agentproto, is a response to a year of protocol sprawl. As the IETF's own blog puts it, "The past year has seen a rush of competing, overlapping protocols for connecting AI agents to one another and to the tools they use—Model Context Protocol (MCP), Agent2Agent (A2A), the Agent Communication Protocol (ACP), the Agent Network Protocol (ANP), and more—alongside a growing number of related Internet-Drafts. The agentproto BoF brings that conversation into the IETF." The session built on prior groundwork: it aimed "to identify which building blocks of agent-to-agent and agent-to-tool communication genuinely need to be standardised, with a view toward chartering a Working Group to take that work forward." That distinction matters. A BoF isn't a vote and it isn't a product launch — the IETF works by rough consensus, and any resulting standard is years away. What's new is the framing: agent-to-agent communication is now being treated as internet-infrastructure territory rather than something each vendor gets to define unilaterally. Two other agent-focused BoFs ran at the same meeting, underscoring how fast this space is moving on the standards side. One, DAWN, is described as tackling "Discovery of Agents, Workloads, and Named Entities, tackling how agents find each other at scale." The other, DMSC, is a diagnostic (non-Working-Group-forming) session mapping gateway-mediated collaboration — how agents expose capabilities, forward requests, and coordinate across trust boundaries. For agentproto specifically, the practical stakes are about who sets the baseline. If a Working Group is chartered, the resulting standard — likely an RFC two to three years out — would establish the protocol baseline that every vendor shipping inter-domain AI agent infrastructure would eventually need to implement. If not, the community returns to the mailing list and the current landscape of competing proprietary-ish standards continues. This is worth flagging now, before the outcome is fully public, because it marks a shift in venue. MCP and A2A have spent the last year maturing inside vendor consortia (Anthropic donated MCP to the Linux Foundation's Agentic AI Foundation; Google donated A2A there too). Moving the interoperability question into an open, rough-consensus IETF process — the same body that standardized TCP/IP, HTTP, and email — is a different kind of commitment than a vendor spec update. It doesn't resolve anything by itself. A charter, if it emerges, still has to survive years of working-group drafting before it becomes an RFC agents actually implement. But it's the first time agent-to-agent coordination has been put in front of the institution that decides what "internet standard" means.