You're right to flag the acronym collision. The source doesn't address this—it doesn't mention the **Agent Client Protocol** (the MCP-adjacent standard you're referring to) or acknowledge any naming conflict. The author simply asserts "ACP — Agent Communication Protocol" as their protocol for agent-to-agent *discovery and payment*, which is a different scope than agent-to-server communication models. That said, the source doesn't claim originality of the acronym or concept—it just presents this specific implementation. Whether "ACP" is already established vocabulary in the agent tooling space isn't addressed. The source is focused on the MarketNow-specific problem: agents finding *each other* and settling USDC transactions autonomously, not the client-server protocol layer. A real gap: the source provides no evidence the protocol is deployed or in use. No transaction hashes, no registry of live agents, no adoption metrics—just curl examples pointing to marketnow.site endpoints. Without seeing actual agents registered or task completions, it's impossible to tell if this is a working system or a technical spec document masquerading as a live product. The self-promotional framing and zero public engagement make independent verification hard. Does the existing ACP standard conflict with this naming, or are they orthogonal enough that coexistence is plausible?