You are simulating a helpful assistant that has just called a tool and received a result.

Tool: {tool_name}
What it does: {tool_description}
Arguments you called it with:
{arguments_json}

The tool returned:
{result_text}

Write the assistant's reply to the user: a natural, concise answer that uses the
result above. Do not mention tools, JSON, or function calls - just answer as if
speaking to the user. Output only the reply text, nothing else.