Autogen

AutoGen 多智能体对话框架助手，精通 Agent 对话编排、代码执行、群聊模式

Security Scan

The skill is an instruction-only helper for the Microsoft AutoGen multi-agent framework and its instructions, dependencies, and capabilities are coherent with that stated purpose.

Purpose & Capability

Name/description claim a multi-agent AutoGen assistant and the SKILL.md contains examples and instructions specific to that framework (agent classes, group chat, tool registration, code execution). Nothing requested (no env vars, no config paths) is unrelated to the stated purpose.

Instruction Scope

Instructions are focused on building and orchestrating agents. They include examples that configure and run code execution (local work_dir and Docker sandbox). This is expected for a framework that executes code, but it does mean the agent flow described can run arbitrary code or start Docker containers if you run the example code.

Install Mechanism

The skill is instruction-only (no install spec, no code files). The README recommends pip installs for upstream packages; the skill itself does not download or install anything. There is no direct install URL or archive baked into the skill.

Credentials

The skill does not declare or require any environment variables or credentials. Example snippets show placing an LLM API key inside an llm_config dict (inline), which is a coding example rather than a demanded secret. Users should avoid embedding production credentials directly in code.

Persistence & Privilege

No always:true, no install-time persistence, and the skill does not request system-level privileges. The skill allows normal autonomous invocation (platform default), which is expected for an agent helper.

Assessment

This skill is coherent with its stated purpose as an AutoGen multi-agent helper. Before using it: (1) Inspect or vet the upstream pip packages it recommends (autogen-agentchat, autogen-ext, pyautogen) — confirm maintainers and package content; (2) If you run the code examples, be aware they can execute arbitrary code and may launch Docker containers — run them in an isolated environment (VM/container) and avoid exposing sensitive host resources; (3) Do not paste production API keys into example code — use secret management or environment variables; (4) If you plan to allow the agent to run autonomously in your environment, consider limiting its privileges (network, filesystem, Docker) and review any code the agent will execute. If you want a deeper check, provide the actual upstream package names/URLs or the code that will be executed so I can look for surprising behavior.

AutoGen 多智能体对话框架助手

你是 Microsoft AutoGen 框架的专家，帮助用户构建基于对话的多智能体协作系统。

核心概念

概念	说明
ConversableAgent	所有 Agent 的基类，支持发送/接收消息、生成回复
AssistantAgent	AI 助手，默认使用 LLM 生成回复
UserProxyAgent	用户代理，可执行代码、请求人类输入
GroupChat	群聊容器，管理多个 Agent 的对话
GroupChatManager	群聊管理器，决定下一个发言的 Agent

安装

pip install autogen-agentchat autogen-ext  # 0.4+ 新架构
pip install pyautogen                       # 经典版本 0.2.x

双人对话模式

from autogen import AssistantAgent, UserProxyAgent

llm_config = {"config_list": [{"model": "gpt-4o", "api_key": "your-key"}]}

assistant = AssistantAgent(
    name="编程助手",
    system_message="你是一个 Python 编程专家。",
    llm_config=llm_config
)

user_proxy = UserProxyAgent(
    name="用户",
    human_input_mode="NEVER",       # NEVER/ALWAYS/TERMINATE
    max_consecutive_auto_reply=10,
    code_execution_config={"work_dir": "coding", "use_docker": False}
)

user_proxy.initiate_chat(assistant, message="写一个快速排序算法并测试性能。")

群聊模式

from autogen import GroupChat, GroupChatManager

planner = AssistantAgent(name="产品经理",
    system_message="你负责分析需求、拆解任务。", llm_config=llm_config)
coder = AssistantAgent(name="开发工程师",
    system_message="你负责编写代码实现功能。", llm_config=llm_config)
reviewer = AssistantAgent(name="代码审查员",
    system_message="你负责审查代码质量。", llm_config=llm_config)
executor = UserProxyAgent(name="执行器",
    human_input_mode="NEVER",
    code_execution_config={"work_dir": "project", "use_docker": True})

group_chat = GroupChat(
    agents=[planner, coder, reviewer, executor],
    messages=[], max_round=20,
    speaker_selection_method="auto"  # auto/round_robin/random/manual
)
manager = GroupChatManager(groupchat=group_chat, llm_config=llm_config)
executor.initiate_chat(manager, message="开发一个 TODO List REST API")

代码执行

# 本地执行
user_proxy = UserProxyAgent(name="executor", code_execution_config={
    "work_dir": "workspace", "use_docker": False, "timeout": 60
})

# Docker 沙箱（推荐生产环境）
user_proxy = UserProxyAgent(name="executor", code_execution_config={
    "work_dir": "workspace", "use_docker": "python:3.11-slim", "timeout": 120
})

工具注册与嵌套对话

from autogen import register_function

def get_weather(city: str) -> str:
    """查询指定城市的天气信息。"""
    return f"{city}：晴，25°C"

register_function(get_weather,
    caller=assistant, executor=user_proxy,
    name="get_weather", description="查询城市天气")

# 嵌套对话：Agent 内部触发子对话
assistant.register_nested_chats(
    trigger=user_proxy,
    chat_queue=[{"recipient": reviewer, "message": "请审查代码。", "max_turns": 3}]
)

与同类框架对比

特性	AutoGen	CrewAI	LangGraph
核心理念	对话驱动协作	角色扮演团队	图状态机
代码执行	原生 Docker 沙箱	需集成工具	需自行实现
群聊模式	原生 GroupChat	不支持	需手动编排
人类介入	human_input_mode 灵活控制	有限支持	中断点机制
嵌套对话	原生支持	不支持	子图实现
适用场景	代码生成、多轮讨论	内容生产、流程自动化	复杂工作流