add-backend-tool

majiayu000 的头像作者 majiayu000

为后端OpenAI函数调用系统添加新工具。当用户提到“新工具”、“添加工具”、“后端功能”、“代理能力”或想要扩展人工智能代理的功能时使用。

📖 中文导读:为后端OpenAI函数调用系统添加新工具。当用户提到“新工具”、“添加工具”、“后端功能”、“代理能力”或想要扩展人工智能代理的功能时使用。 以下正文为该技能的原始 SKILL.md 内容(英文),上方安装命令可直接使用。

name: add-backend-tool description: Add a new tool to the backend OpenAI function calling system. Use when user mentions "new tool", "add tool", "backend function", "agent capability", or wants to extend what the AI agent can do.

Add Backend Tool

Instructions

  1. Read backend/main.py to understand existing tool patterns:

    • Find the tools list with function definitions
    • Review helper functions (read_file, write_file, run_terminal_command, web_search)
  2. Create the helper function:

    def new_tool_name(param1: str, param2: int = 10) -> str:
        """Docstring explaining the tool."""
        try:
            # Implementation
            return result
        except Exception as e:
            return f"Error: {str(e)}"
    
  3. Add tool definition to the tools list:

    {
        "type": "function",
        "function": {
            "name": "new_tool_name",
            "description": "What this tool does and when to use it",
            "parameters": {
                "type": "object",
                "properties": {
                    "param1": {"type": "string", "description": "..."},
                    "param2": {"type": "integer", "description": "..."}
                },
                "required": ["param1"]
            }
        }
    }
    
  4. Add tool invocation handler in the WebSocket message loop:

    elif func_name == "new_tool_name":
        result = new_tool_name(**args)
    
  5. Update mode restrictions if needed:

    • Agent mode: full access
    • Chat mode: add to allowed list only if read-only/safe

Examples

  • "Add a tool to list directory contents"
  • "Create a tool for git operations"
  • "Add web scraping capability"

Guardrails

  • Include proper error handling with try/except
  • Add timeout for any long-running operations
  • Consider security implications (Chat mode restrictions)
  • Never hardcode API keys or secrets
  • Document the tool's purpose in the function docstring