docs(ai): list connect-mcp in the AI module landing index

content/docs/ai/meta.json ships 9 pages; the "What's in this module" list
in index.mdx linked 7 of the 8 content pages. connect-mcp was absent and
the string appeared nowhere else on the page, so the module's only
inventory did not reach meta.json's first content page.

Placement is derived, not chosen: the existing 7 were already exactly
meta.json's order minus the missing one, so the row goes first, at its
meta.json position, rather than being appended.

Link text is the page's own frontmatter title ("Connect an MCP Client"),
not a slug prettification. All 7 existing rows carry an em-dash gloss, so
the new row carries one too.

Also records meta.json as the source of truth in a comment beside the
list, mirroring the ruling already applied to permissions/index.mdx.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DdCnBGcHeufjrq7drTD3wt

Author: claude

content/docs/ai/index.mdx

@@ -25,6 +25,11 @@ AI in ObjectStack is a **cross-protocol capability layer**: agents, tools, and k
 
 ## What's in this module
 
+{/* Source of truth for this module's page set AND its order is meta.json in this directory — routing and the sidebar are built from it, and the module rows below mirror it exactly, minus this page. The two trailing rows (Spec / Schema reference) link out of the module and have no meta.json entry. */}
+{/* The list is hand-kept on purpose: meta.json stores slugs only, so the link text and the glosses below exist nowhere else, and a generated list would lose them. */}
+{/* Drift note: nothing checks the two against each other in either direction, so a page added to or removed from meta.json leaves this list silently stale — edit both in the same change. */}
+
+- [Connect an MCP Client](/docs/ai/connect-mcp) — point Claude Code, Claude Desktop, or any MCP client at the `/api/v1/mcp` server every deployment already runs, over OAuth or an API key
 - [AI Agents](/docs/ai/agents) — the two platform agents (`ask` / `build`), skills as the extension primitive, and agent anatomy
 - [Actions as Tools](/docs/ai/actions-as-tools) — explicit opt-in exposure of Actions to the LLM, HITL approval, permission-aware execution
 - [Tool Records](/docs/ai/tools) — the three ways a capability reaches an agent, and the narrow case where authoring a `tool` record is the right answer