docs(integrations): add Pydantic AI FastMCP toolset guide

[MukundaKatta]

Why

Closes #2246. Pydantic AI ships a FastMCPToolset (in pydantic_ai.toolsets.fastmcp) that lets a Pydantic AI agent call tools from any FastMCP (or other MCP) server, but the FastMCP docs do not cover the integration. New users searching the FastMCP docs for Pydantic AI find nothing and have to dig through the Pydantic AI repo to wire it up.

What

• New page at docs/integrations/pydantic-ai.mdx under the existing AI SDKs group, matching the style of the Anthropic, OpenAI, and Gemini integration pages.
• Walks through pointing FastMCPToolset at a FastMCP server with examples for in-memory, Streamable HTTP, STDIO, and mcpServers JSON config transports.
• Covers auth pass-through by building a Client / StreamableHttpTransport with a bearer header, and links to the FastMCP OAuth helper plus server-side Token Verification.
• Links out to the upstream Pydantic AI FastMCPToolset docs so readers can find the full API.
• Adds the new page to docs/docs.json so it shows up in the AI SDKs nav alongside Anthropic / Gemini / OpenAI.

Tested

• Verified FastMCPToolset import path and accepted constructor arg types against the Pydantic AI source (pydantic_ai_slim/pydantic_ai/toolsets/fastmcp.py) and Pydantic AI's docs/mcp/fastmcp-client.md.
• Verified the new entry slots into docs/docs.json under the existing AI SDKs group without breaking JSON structure.
• Confirmed all internal doc links (/clients/client, /clients/transports, /clients/auth, /clients/auth/oauth, /integrations/mcp-json-configuration, /servers/auth/token-verification) resolve to existing pages in the repo.

[claude]

Claude Code Review

This pull request is from a fork — automated review is disabled. A repository maintainer can comment @claude review to run a one-time review.

[jlowin]

Thanks for the contribution! Solid integration guide. I'll merge as-is and follow up with a tiny PR to fix two broken doc-link anchors and hoist an import — minor enough to not block on a re-push.