AgentRFC: Security Design Principles and Conformance Testing for Agent Protocols

AgentRFC is a research contribution (arxiv paper, 13 days old, 0 stars) that addresses a genuine gap: systematic security design for rapidly-deployed agent protocols. The framing as a 6-layer stack (analogous to ITU-T X.800) is conceptually sound and reusable. However, defensibility is threatened on multiple fronts: (1) PLATFORM DOMINATION: OpenAI (MCP owner), Anthropic, Google, and Microsoft are all actively building/standardizing agent protocol security. They have resources to incorporate RFC-style guidance into native implementations and SDKs. The fact that MCP has 97M+ monthly downloads means the protocol owners themselves will likely drive standardization rather than external researchers. (2) MARKET CONSOLIDATION: Protocol security is not a separate market—it's absorbed into protocol governance. The protocol consortiums (Anthropic for MCP, others for A2A/ANP/ACP) will adopt/fork/ignore these recommendations based on strategic fit, not open-source momentum. (3) TIMING: The paper is a theoretical framework + testing methodology. Once published, it becomes a reference standard, but implementation/adoption depends on protocol owners' roadmaps, not community forks (only 2 so far). The 13-day age and 0 velocity suggest this is pre-launch. DEFENSIBILITY RATIONALE: Score 7 reflects that this is a well-positioned reference framework (strong technical contribution, timely topic, addresses real pain point), but it lacks a composable implementation moat. It will be cited and potentially inform standards, but will not become a standalone competitive product. The security analysis itself is defensible as intellectual property / standards contribution, but cannot be 'displaced' in the traditional sense—it will be adopted, forked, or ignored based on protocol governance decisions, not technical merit alone. RISK TIMELINE: Platforms and protocol consortiums are actively defining security postures for these protocols. Within 1-2 years, we expect MCP v2+ and other protocols to integrate RFC-style security layers into their specs. The window for external standard-setting is closing rapidly. If the authors can get this adopted by the MCP SC or Anthropic's governance process, defensibility jumps to 8-9. Otherwise, it remains a valuable but non-moated reference.