refactor: use ToolResult.get_serialized_output() for context messages

Use the new get_serialized_output() method from amplifier-core to
properly serialize tool results for LLM context. This ensures structured
outputs (like bash tool's stdout/stderr/returncode dict) are JSON-
serialized rather than using Python repr format.

🤖 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

@@ -310,10 +310,8 @@ async def execute_single_tool(tc: Any, group_id: str) -> tuple[str, str]:
                             if coordinator:
                                 await coordinator.process_hook_result(post_result, "tool:post", tool_name)
 
-                            # Return success with result content
-                            result_content = str(
-                                getattr(result, "data", None) or getattr(result, "text", None) or result
-                            )
+                            # Return success with result content (JSON-serialized for dict/list)
+                            result_content = result.get_serialized_output()
                             return (tool_call_id, result_content)
 
                         except Exception as te: