fix: Remove authorization from provider data (#4161)

# What does this PR do?
- Remove backward compatibility for authorization in mcp_headers
- Enforce authorization must use dedicated parameter  
- Add validation error if Authorization found in provider_data headers
- Update test_mcp.py to use authorization parameter
- Update test_mcp_json_schema.py to use authorization parameter
- Update test_tools_with_schemas.py to use authorization parameter
- Update documentation to show the change in the authorization approach

Breaking Change:
- Authorization can no longer be passed via mcp_headers in provider_data
- Users must use the dedicated 'authorization' parameter instead
- Clear error message guides users to the new approach"

## Test Plan
CI

---------

Co-authored-by: Omar Abdelwahab <omara@fb.com>
Co-authored-by: Ashwin Bharambe <ashwin.bharambe@gmail.com>

tests/integration/inference/test_tools_with_schemas.py

@@ -9,8 +9,6 @@ Integration tests for inference/chat completion with JSON Schema-based tools.
 Tests that tools pass through correctly to various LLM providers.
 """
 
-import json
-
 import pytest
 
 from llama_stack.core.library_client import LlamaStackAsLibraryClient
@@ -193,22 +191,11 @@ class TestMCPToolsInChatCompletion:
             mcp_endpoint=dict(uri=uri),
         )
 
-        # Use old header-based approach for Phase 1 (backward compatibility)
-        provider_data = {
-            "mcp_headers": {
-                uri: {
-                    "Authorization": f"Bearer {AUTH_TOKEN}",
-                },
-            },
-        }
-        auth_headers = {
-            "X-LlamaStack-Provider-Data": json.dumps(provider_data),
-        }
-
+        # Use the dedicated authorization parameter
         # Get the tools from MCP
         tools_response = llama_stack_client.tool_runtime.list_tools(
             tool_group_id=test_toolgroup_id,
-            extra_headers=auth_headers,
+            authorization=AUTH_TOKEN,
         )
 
         # Convert to OpenAI format for inference