人机交互 (HITL)

某些工具操作可能是敏感的，需要在执行前获得人工批准。Deep agents 通过 LangGraph 的中断功能支持人机交互工作流。您可以使用 interrupt_on 参数配置哪些工具需要批准。

基础配置

interrupt_on 参数接受一个将工具名称映射到中断配置的字典。每个工具可以配置为：

True：启用具有默认行为的中断（允许批准、编辑、拒绝）
False：对此工具禁用中断
{"allowed_decisions": [...]}：具有特定允许决策的自定义配置

from langchain.tools import tool
from deepagents import create_deep_agent
from langgraph.checkpoint.memory import MemorySaver

@tool
def delete_file(path: str) -> str:
    """Delete a file from the filesystem."""
    return f"Deleted {path}"

@tool
def read_file(path: str) -> str:
    """Read a file from the filesystem."""
    return f"Contents of {path}"

@tool
def send_email(to: str, subject: str, body: str) -> str:
    """Send an email."""
    return f"Sent email to {to}"

# Checkpointer is REQUIRED for human-in-the-loop
checkpointer = MemorySaver()

agent = create_deep_agent(
    model="claude-sonnet-4-6",
    tools=[delete_file, read_file, send_email],
    interrupt_on={
        "delete_file": True,  # Default: approve, edit, reject
        "read_file": False,   # No interrupts needed
        "send_email": {"allowed_decisions": ["approve", "reject"]},  # No editing
    },
    checkpointer=checkpointer  # Required!
)

决策类型

allowed_decisions 列表控制人工在审查工具调用时可以采取的操作：

"approve"：按 agent 提议的原样执行工具和参数
"edit"：在执行前修改工具参数
"reject"：完全跳过执行此工具调用

您可以自定义每个工具可用的决策：

const interruptOn = {
  // Sensitive operations: allow all options
  delete_file: { allowedDecisions: ["approve", "edit", "reject"] },

  // Moderate risk: approval or rejection only
  write_file: { allowedDecisions: ["approve", "reject"] },

  // Must approve (no rejection allowed)
  critical_operation: { allowedDecisions: ["approve"] },
};

处理中断

当触发中断时，agent 暂停执行并返回控制权。检查结果中的中断并相应地处理它们。

import { v4 as uuidv4 } from "uuid";
import { Command } from "@langchain/langgraph";

// Create config with thread_id for state persistence
const config = { configurable: { thread_id: uuidv4() } };

// Invoke the agent
let result = await agent.invoke({
  messages: [{ role: "user", content: "Delete the file temp.txt" }],
}, config);

// Check if execution was interrupted
if (result.__interrupt__) {
  // Extract interrupt information
  const interrupts = result.__interrupt__[0].value;
  const actionRequests = interrupts.actionRequests;
  const reviewConfigs = interrupts.reviewConfigs;

  // Create a lookup map from tool name to review config
  const configMap = Object.fromEntries(
    reviewConfigs.map((cfg) => [cfg.actionName, cfg])
  );

  // Display the pending actions to the user
  for (const action of actionRequests) {
    const reviewConfig = configMap[action.name];
    console.log(`Tool: ${action.name}`);
    console.log(`Arguments: ${JSON.stringify(action.args)}`);
    console.log(`Allowed decisions: ${reviewConfig.allowedDecisions}`);
  }

  // Get user decisions (one per actionRequest, in order)
  const decisions = [
    { type: "approve" }  // User approved the deletion
  ];

  // Resume execution with decisions
  result = await agent.invoke(
    new Command({ resume: { decisions } }),
    config  // Must use the same config!
  );
}

// Process final result
console.log(result.messages[result.messages.length - 1].content);

多个工具调用

当 agent 调用多个需要批准的工具时，所有中断都将在单个中断中批处理在一起。您必须按顺序为每个工具提供决策。

const config = { configurable: { thread_id: uuidv4() } };

let result = await agent.invoke({
  messages: [{
    role: "user",
    content: "Delete temp.txt and send an email to admin@example.com"
  }]
}, config);

if (result.__interrupt__) {
  const interrupts = result.__interrupt__[0].value;
  const actionRequests = interrupts.actionRequests;

  // Two tools need approval
  console.assert(actionRequests.length === 2);

  // Provide decisions in the same order as actionRequests
  const decisions = [
    { type: "approve" },  // First tool: delete_file
    { type: "reject" }    // Second tool: send_email
  ];

  result = await agent.invoke(
    new Command({ resume: { decisions } }),
    config
  );
}

编辑工具参数

当 "edit" 在允许的决策中时，您可以在执行前修改工具参数：

if (result.__interrupt__) {
  const interrupts = result.__interrupt__[0].value;
  const actionRequest = interrupts.actionRequests[0];

  // Original args from the agent
  console.log(actionRequest.args);  // { to: "everyone@company.com", ... }

  // User decides to edit the recipient
  const decisions = [{
    type: "edit",
    editedAction: {
      name: actionRequest.name,  // Must include the tool name
      args: { to: "team@company.com", subject: "...", body: "..." }
    }
  }];

  result = await agent.invoke(
    new Command({ resume: { decisions } }),
    config
  );
}

子 agent 中断

当使用子 agent 时，您可以在工具调用上和在工具调用内使用中断。

工具调用上的中断

每个子 agent 可以有自己的 interrupt_on 配置，该配置会覆盖主 agent 的设置：

const agent = createDeepAgent({
  tools: [deleteFile, readFile],
  interruptOn: {
    delete_file: true,
    read_file: false,
  },
  subagents: [{
    name: "file-manager",
    description: "Manages file operations",
    systemPrompt: "You are a file management assistant.",
    tools: [deleteFile, readFile],
    interruptOn: {
      // Override: require approval for reads in this subagent
      delete_file: true,
      read_file: true,  // Different from main agent!
    }
  }],
  checkpointer
});

当子 agent 触发中断时，处理方式是相同的——检查 __interrupt__ 并使用 Command 恢复。

工具调用内的中断

子 agent 工具可以直接调用 interrupt() 来暂停执行并等待批准：

import { createAgent, tool } from "langchain";
import { ChatOpenAI } from "@langchain/openai";
import { HumanMessage } from "@langchain/core/messages";
import { MemorySaver, Command, interrupt } from "@langchain/langgraph";
import { createDeepAgent } from "deepagents";
import { z } from "zod";

const requestApproval = tool(
  async ({ actionDescription }: { actionDescription: string }) => {
    const approval = interrupt({
      type: "approval_request",
      action: actionDescription,
      message: `Please approve or reject: ${actionDescription}`,
    }) as { approved?: boolean; reason?: string };

    if (approval.approved) {
      return `Action '${actionDescription}' was APPROVED. Proceeding...`;
    } else {
      return `Action '${actionDescription}' was REJECTED. Reason: ${
        approval.reason || "No reason provided"
      }`;
    }
  },
  {
    name: "request_approval",
    description: "Request human approval before proceeding with an action.",
    schema: z.object({
      actionDescription: z
        .string()
        .describe("The action that requires approval"),
    }),
  }
);

async function main() {
  const checkpointer = new MemorySaver();
  const model = new ChatOpenAI({
    model: "gpt-4o-mini",
    maxTokens: 4096,
  });

  const compiledSubagent = createAgent({
    model: model,
    tools: [requestApproval],
    name: "approval-agent",
  });

  const parentAgent = await createDeepAgent({
    checkpointer: checkpointer,
    subagents: [
      {
        name: "approval-agent",
        description: "An agent that can request approvals",
        runnable: compiledSubagent as any,
      },
    ],
  });

  const threadId = "test_interrupt_directly";
  const config = { configurable: { thread_id: threadId } };

  console.log("Invoking agent - sub-agent will use request_approval tool...");

  let result = await parentAgent.invoke(
    {
      messages: [
        new HumanMessage({
          content:
            "Use the task tool to launch the approval-agent sub-agent. " +
            "Tell it to use the request_approval tool to request approval for 'deploying to production'.",
        }),
      ],
    },
    config
  );

  if (result.__interrupt__) {
    const interruptValue = result.__interrupt__[0].value as {
      type?: string;
      action?: string;
      message?: string;
    };
    console.log("\nInterrupt received!");
    console.log(`  Type: ${interruptValue.type}`);
    console.log(`  Action: ${interruptValue.action}`);
    console.log(`  Message: ${interruptValue.message}`);

    console.log("\nResuming with Command(resume={'approved': true})...");
    const result2 = await parentAgent.invoke(
      new Command({ resume: { approved: true } }),
      config
    );

    if (!result2.__interrupt__) {
      console.log("\nExecution completed!");
      // Find the tool response
      const toolMsgs = result2.messages?.filter((m) => m.type === "tool") || [];
      if (toolMsgs.length > 0) {
        const lastToolMsg = toolMsgs[toolMsgs.length - 1];
        console.log(`  Tool result: ${lastToolMsg.content}`);
      }
    } else {
      console.log("\nAnother interrupt occurred");
    }
  } else {
    console.log(
      "\n  No interrupt - the model may not have called request_approval"
    );
  }
}

main().catch(console.error);

运行时，这会产生以下输出：

Invoking agent - sub-agent will use request_approval tool...

Interrupt received!
  Type: approval_request
  Action: deploying to production
  Message: Please approve or reject: deploying to production

Resuming with Command(resume={'approved': true})...

Execution completed!
  Tool result: Approval for "deploying to production" has been granted. You can proceed with the deployment.

最佳实践

始终使用检查点

人机交互需要在中断和恢复之间持久化 agent 状态：

使用相同的线程 ID

恢复时，您必须使用具有相同 thread_id 的相同配置：

决策顺序与动作匹配

决策列表必须与 action_requests 的顺序匹配：

按风险定制配置

根据风险级别配置不同的工具：

在 GitHub 上编辑此页面或提交 issue.

将这些文档连接到 Claude、VSCode 以及更多通过 MCP 获取实时答案的工具。

Get started

Core capabilities

Streaming

Deep Agents CLI

基础配置

决策类型

处理中断

多个工具调用

编辑工具参数

子 agent 中断

工具调用上的中断

工具调用内的中断

最佳实践

始终使用检查点

使用相同的线程 ID

决策顺序与动作匹配

按风险定制配置

Get started

Core capabilities

Streaming

Deep Agents CLI

​基础配置

​决策类型

​处理中断

​多个工具调用

​编辑工具参数

​子 agent 中断

​工具调用上的中断

​工具调用内的中断

​最佳实践

​始终使用检查点

​使用相同的线程 ID

​决策顺序与动作匹配

​按风险定制配置

基础配置

决策类型

处理中断

多个工具调用

编辑工具参数

子 agent 中断

工具调用上的中断

工具调用内的中断

最佳实践

始终使用检查点

使用相同的线程 ID

决策顺序与动作匹配

按风险定制配置