GuidesAgents

用 LangChain Supervisor 模式构建个人助理 Multi-Agent

按官方 subagents personal assistant 教程：日历与邮件子代理各自专注，再包装成工具，由 supervisor 协调跨域请求。

基于 LangChain Docs 整理 · 官方资料 ↗

一个 Agent 同时管日历 API 和邮件 API 时，工具太多、格式太碎，路由容易乱。Supervisor 模式把专家拆成子代理，由中心 supervisor 只看高层工具并协调结果。

这篇跟官方「Build a personal assistant with subagents」：日历助手 + 邮件助手，再合成个人助理。工具在教程里是 stub，重点是分层结构。

为什么要用 Supervisor

多 Agent 把工具与提示按域切开。日历只懂排期，邮件只懂写信；supervisor 决定何时 schedule_event、何时 manage_email，而不是在几十个底层 API 里挑。

适合：多个独立域、子代理不必直接对用户说话、你希望工作流中心化控制。

先定义域内 Tools

官方安装

bash

pip install langchain

官方 stub tools

python

from langchain.tools import tool

@tool
def create_calendar_event(
    title: str,
    start_time: str,
    end_time: str,
    attendees: list[str],
    location: str = "",
) -> str:
    """Create a calendar event. Requires exact ISO datetime format."""
    return f"Event created: {title} from {start_time} to {end_time} with {len(attendees)} attendees"

@tool
def send_email(
    to: list[str],
    subject: str,
    body: str,
    cc: list[str] = [],
) -> str:
    """Send an email via email API. Requires properly formatted addresses."""
    return f"Email sent to {', '.join(to)} - Subject: {subject}"

@tool
def get_available_time_slots(
    attendees: list[str],
    date: str,
    duration_minutes: int,
) -> list[str]:
    """Check calendar availability for given attendees on a specific date."""
    return ["09:00", "14:00", "16:00"]

做成专注的子代理

日历与邮件子代理

python

from datetime import date
from langchain.agents import create_agent

CALENDAR_AGENT_PROMPT = (
    f"Today's date is {date.today().isoformat()}. "
    "You are a calendar scheduling assistant. "
    "Parse natural language scheduling requests into proper ISO datetime formats. "
    "Use get_available_time_slots to check availability when needed. "
    "Use create_calendar_event to schedule events. "
    "Always confirm what was scheduled in your final response."
)

calendar_agent = create_agent(
    model,
    tools=[create_calendar_event, get_available_time_slots],
    system_prompt=CALENDAR_AGENT_PROMPT,
)

EMAIL_AGENT_PROMPT = (
    "You are an email assistant. "
    "Compose professional emails based on natural language requests. "
    "Use send_email to send the message. "
    "Always confirm what was sent in your final response."
)

email_agent = create_agent(
    model,
    tools=[send_email],
    system_prompt=EMAIL_AGENT_PROMPT,
)

包装成工具，再交给 Supervisor

关键一步：子代理对 supervisor 只暴露高层工具。supervisor 看到的是 schedule_event / manage_email，不是 create_calendar_event。

子代理包装 + supervisor

python

@tool
def schedule_event(request: str) -> str:
    """Schedule calendar events using natural language."""
    result = calendar_agent.invoke({
        "messages": [{"role": "user", "content": request}]
    })
    return result["messages"][-1].text

@tool
def manage_email(request: str) -> str:
    """Send emails using natural language."""
    result = email_agent.invoke({
        "messages": [{"role": "user", "content": request}]
    })
    return result["messages"][-1].text

SUPERVISOR_PROMPT = (
    "You are a helpful personal assistant. "
    "You can schedule calendar events and send emails. "
    "Break down user requests into appropriate tool calls and coordinate the results. "
    "When a request involves multiple actions, use multiple tools in sequence or in parallel as appropriate."
)

supervisor_agent = create_agent(
    model,
    tools=[schedule_event, manage_email],
    system_prompt=SUPERVISOR_PROMPT,
)

跨域请求示例

python

query = "Schedule a team standup for tomorrow at 9am"
# 或同时排会 + 发提醒的组合请求

result = supervisor_agent.invoke({
    "messages": [{"role": "user", "content": query}]
})
print(result["messages"][-1].text)

返回给 supervisor 的 ideally 只有子代理最终回复，避免把中间 tool 轨迹灌进上层上下文。官方后续还可加 human-in-the-loop 审核外发邮件。

容易踩的坑

supervisor 仍直接调用底层 API 工具

确认只把 schedule_event / manage_email 传给 supervisor，底层 tools 只挂在子代理上。

子代理上下文泄漏到上层

包装工具只返回 messages[-1].text，不要把完整中间轨迹回传。

还在找 langgraph-supervisor 包

当前成熟教程已迁到 LangChain multi-agent subagents 模式；迁移说明见官方 Migrate from langgraph-supervisor。

官方资料

LangChain Docs

Build a personal assistant with subagents ↗