FullScope-MCP

Instructions

调用模型进行回答

Args:
    prompt: 要发送给模型的提示词

Returns:
    模型的回答

Implementation Reference

• src/fullscope_mcp_server/server.py:277-302 (handler)

  The `call_model` tool handler function, registered via `@mcp.tool()`. It calls the OpenAI chat completions API using the `summarizer.client` with the provided `prompt`, handling errors gracefully.

  @mcp.tool()
  async def call_model(prompt: str, ctx: Context) -> str:
      """
      调用模型进行回答
      
      Args:
          prompt: 要发送给模型的提示词
      
      Returns:
          模型的回答
      """
      try:
          response = summarizer.client.chat.completions.create(
              model=OPENAI_MODEL,
              messages=[
                  {"role": "user", "content": prompt}
              ],
              max_tokens=MAX_OUTPUT_TOKENS,
              temperature=0.7
          )
          
          return response.choices[0].message.content.strip()
          
      except Exception as e:
          logger.error(f"模型调用失败: {e}")
          return f"模型调用失败: {str(e)}"

• src/fullscope_mcp_server/server.py:277-277 (registration)

  The `@mcp.tool()` decorator registers the `call_model` function as an MCP tool.

  @mcp.tool()

• src/fullscope_mcp_server/server.py:278-287 (schema)

  Function signature and docstring define the input schema (prompt: str, ctx: Context) and output (str).

  async def call_model(prompt: str, ctx: Context) -> str:
      """
      调用模型进行回答
      
      Args:
          prompt: 要发送给模型的提示词
      
      Returns:
          模型的回答
      """