feat: pass provider to context.get_messages_for_request()

Enable dynamic token budget calculation by passing provider reference
to context manager's get_messages_for_request() method.

🤖 Generated with [Amplifier](https://github.com/microsoft/amplifier)

Co-Authored-By: Amplifier <240397093+microsoft-amplifier@users.noreply.github.com>

Author: bkrabach

amplifier_module_loop_basic/__init__.py

@@ -92,8 +92,9 @@ async def execute(
                     return f"Operation denied: {result.reason}"
 
             # Get messages for LLM request (context handles compaction internally)
+            # Pass provider for dynamic budget calculation based on model's context window
             if hasattr(context, "get_messages_for_request"):
-                message_dicts = await context.get_messages_for_request()
+                message_dicts = await context.get_messages_for_request(provider=provider)
             else:
                 # Fallback for simple contexts without the method
                 message_dicts = getattr(context, "messages", [{"role": "user", "content": prompt}])
@@ -412,7 +413,7 @@ async def execute_single_tool(tc: Any, group_id: str) -> tuple[str, str]:
 
             # Get one final response with the reminder (context handles compaction internally)
             if hasattr(context, "get_messages_for_request"):
-                message_dicts = await context.get_messages_for_request()
+                message_dicts = await context.get_messages_for_request(provider=provider)
             else:
                 message_dicts = getattr(context, "messages", [{"role": "user", "content": prompt}])
             message_dicts = list(message_dicts)