MCP server with SSE example not implemented properly.

**Describe the bug**

The documentation section titled **"Using MCP Tools in your own Agent out of adk web"** claims to demonstrate how to use MCP tools in agents outside of the `adk-web` UI, particularly when connecting to a remote MCP server via SSE. However, the code example still uses `StdioServerParameters`, which is already covered in earlier examples and does not match the intended use case described in the section.

**To Reproduce**

Steps to reproduce the behavior:

1. Go to [https://google.github.io/adk-docs/tools/mcp-tools/#using-mcp-tools-in-your-own-agent-out-of-adk-web](https://google.github.io/adk-docs/tools/mcp-tools/#using-mcp-tools-in-your-own-agent-out-of-adk-web)
2. Read the introductory text about connecting to remote MCP servers or running tools in another process.
3. Observe that the example still uses `StdioServerParameters`, not `SseServerParams`.

**Expected behavior**

A working example using `SseServerParams` to connect to a remote or separately running MCP server, such as:

```python
toolset = MCPToolset(
connection_params=SseServerParams(
url="http://localhost:3000/mcp",
headers={"Authorization": "Bearer your-token"} # if required
),
tool_filter=["read_file", "list_directory"]
)
```

This would more accurately reflect the section’s purpose and be helpful to developers integrating ADK into their own UIs or non-web apps.

**Screenshots**

N/A

**Versions**

* OS: \[e.g. Windows, Mac, Linux]
* ADK version: \[unspecified]
* Python version: \[unspecified]

**Additional context**

This distinction is important for users who are building custom agent experiences outside of `adk-web` and need clear examples for integrating with remotely hosted MCP tools via SSE.