Full Answer

The seven protocols break into three categories. Identity files — agents.txt and agent-card.json — tell an arriving agent who the site is and what it offers. Capability manifests — mcp.json, ai-catalog.json, and openapi.json — describe what the agent can do once connected. Discovery conventions — llms.txt and .well-known entries — help the agent find the other files in the first place. No single standard covers all three categories, which is why adoption has stalled.

The fragmentation creates a real problem for store owners. An AI shopping agent built on one protocol cannot discover a store exposing another. A site that implements agents.txt but not openapi.json is visible but not transactable. A site with openapi.json but no identity file is transactable but anonymous — the agent cannot verify it is dealing with the business it intended to reach.

For WordPress and WooCommerce sites, the REST API already provides the machinery. A custom namespace can register rooms for identity, catalogue, pricing, transactions, and service — all behind native authentication. WooCommerce's own [MCP server integration](https://seresa.io/blog/ai-data-readiness/wordpress-mcp-servers-explained) already handles one protocol natively. The missing piece is the [side door](https://seresa.io/blog/side-door-ai-agents-polymath-website) that connects those rooms to whichever discovery protocol an agent uses, so the store is ready for all seven without betting on one.