from langchain.agents import create_agent
from langchain.agents.middleware import SummarizationMiddleware

agent = create_agent(
    model="gpt-5.5",
    tools=[your_weather_tool, your_calculator_tool],
    middleware=[
        SummarizationMiddleware(
            model="gpt-5.4-mini",
            trigger=("tokens", 4000),
            keep=("messages", 20),
        ),
    ],
)

from langchain.agents import create_agent
from langchain.agents.middleware import HumanInTheLoopMiddleware
from langgraph.checkpoint.memory import InMemorySaver


def your_read_email_tool(email_id: str) -> str:
    """用于通过 ID 读取电子邮件的模拟函数。"""
    return f"Email content for ID: {email_id}"

def your_send_email_tool(recipient: str, subject: str, body: str) -> str:
    """用于发送电子邮件的模拟函数。"""
    return f"Email sent to {recipient} with subject '{subject}'"

agent = create_agent(
    model="gpt-5.5",
    tools=[your_read_email_tool, your_send_email_tool],
    checkpointer=InMemorySaver(),
    middleware=[
        HumanInTheLoopMiddleware(
            interrupt_on={
                "your_send_email_tool": {
                    "allowed_decisions": ["approve", "edit", "reject"],
                },
                "your_read_email_tool": False,
            }
        ),
    ],
)

from langchain.agents import create_agent
from langchain.agents.middleware import ModelCallLimitMiddleware
from langgraph.checkpoint.memory import InMemorySaver

agent = create_agent(
    model="gpt-5.5",
    checkpointer=InMemorySaver(),  # 线程限制所需
    tools=[],
    middleware=[
        ModelCallLimitMiddleware(
            thread_limit=10,
            run_limit=5,
            exit_behavior="end",
        ),
    ],
)

from langchain.agents import create_agent
from langchain.agents.middleware import ToolCallLimitMiddleware

agent = create_agent(
    model="gpt-5.5",
    tools=[search_tool, database_tool],
    middleware=[
        # 全局限制
        ToolCallLimitMiddleware(thread_limit=20, run_limit=10),
        # 特定工具限制
        ToolCallLimitMiddleware(
            tool_name="search",
            thread_limit=5,
            run_limit=3,
        ),
    ],
)

from langchain.agents import create_agent
from langchain.agents.middleware import ModelFallbackMiddleware

agent = create_agent(
    model="gpt-5.5",
    tools=[],
    middleware=[
        ModelFallbackMiddleware(
            "gpt-5.4-mini",
            "claude-3-5-sonnet-20241022",
        ),
    ],
)

from langchain.agents import create_agent
from langchain.agents.middleware import PIIMiddleware

agent = create_agent(
    model="gpt-5.5",
    tools=[],
    middleware=[
        PIIMiddleware("email", strategy="redact", apply_to_input=True),
        PIIMiddleware("credit_card", strategy="mask", apply_to_input=True),
    ],
)

from langchain.agents import create_agent
from langchain.agents.middleware import PIIMiddleware
import re


# 方法 1: 正则表达式模式字符串
agent1 = create_agent(
    model="gpt-5.5",
    tools=[],
    middleware=[
        PIIMiddleware(
            "api_key",
            detector=r"sk-[a-zA-Z0-9]{32}",
            strategy="block",
        ),
    ],
)

# 方法 2: 编译后的正则表达式模式
agent2 = create_agent(
    model="gpt-5.5",
    tools=[],
    middleware=[
        PIIMiddleware(
            "phone_number",
            detector=re.compile(r"\+?\d{1,3}[\s.-]?\d{3,4}[\s.-]?\d{4}"),
            strategy="mask",
        ),
    ],
)

# 方法 3: 自定义检测器函数
def detect_ssn(content: str) -> list[dict[str, str | int]]:
    """检测 SSN 并进行验证。

    返回一个字典列表，包含 'text'、'start' 和 'end' 键。
    """
    import re
    matches = []
    pattern = r"\d{3}-\d{2}-\d{4}"
    for match in re.finditer(pattern, content):
        ssn = match.group(0)
        # 验证：前 3 位数字不应为 000、666 或 900-999
        first_three = int(ssn[:3])
        if first_three not in [0, 666] and not (900 <= first_three <= 999):
            matches.append({
                "text": ssn,
                "start": match.start(),
                "end": match.end(),
            })
    return matches

agent3 = create_agent(
    model="gpt-5.5",
    tools=[],
    middleware=[
        PIIMiddleware(
            "ssn",
            detector=detect_ssn,
            strategy="hash",
        ),
    ],
)

def detector(content: str) -> list[dict[str, str | int]]:
    return [
        {"text": "matched_text", "start": 0, "end": 12},
        # ... 更多匹配项
    ]

from langchain.agents import create_agent
from langchain.agents.middleware import TodoListMiddleware

agent = create_agent(
    model="gpt-5.5",
    tools=[read_file, write_file, run_tests],
    middleware=[TodoListMiddleware()],
)

from langchain.agents import create_agent
from langchain.agents.middleware import LLMToolSelectorMiddleware

agent = create_agent(
    model="gpt-5.5",
    tools=[tool1, tool2, tool3, tool4, tool5, ...],
    middleware=[
        LLMToolSelectorMiddleware(
            model="gpt-5.4-mini",
            max_tools=3,
            always_include=["search"],
        ),
    ],
)

from langchain.agents import create_agent
from langchain.agents.middleware import ToolErrorMiddleware


def on_error(exc: Exception, request: ToolCallRequest) -> str | None:
    if isinstance(exc, ValueError):
        return f"`{request.tool_call['name']}` failed with {type(exc).__name__}."
    # 传播其他所有异常


agent = create_agent(
    model="gpt-5.5",
    tools=[your_tools],
    middleware=[ToolErrorMiddleware(on_error)],
)

<Note>
    建议返回包含异常类型名称的内容，而不是原始的异常消息，因为后者可能包含敏感或内部细节。`on_error` 处理程序控制信息的披露：原始异常消息永远不会发送给模型，除非你选择包含它。
</Note>

</Accordion>
:::

### 工具重试

自动重试失败的工具调用，支持可配置的指数退避。工具重试在以下场景中很有用：

- 处理外部 API 调用中的瞬态故障。
- 提高依赖网络的工具的可靠性。
- 构建能优雅处理临时错误的弹性代理。

**API 参考：** @[`ToolRetryMiddleware`]

```python
from langchain.agents import create_agent
from langchain.agents.middleware import ToolRetryMiddleware

agent = create_agent(
    model="gpt-5.5",
    tools=[search_tool, database_tool],
    middleware=[
        ToolRetryMiddleware(
            max_retries=3,
            backoff_factor=2.0,
            initial_delay=1.0,
        ),
    ],
)

from langchain.agents import create_agent
from langchain.agents.middleware import ModelRetryMiddleware

agent = create_agent(
    model="gpt-5.5",
    tools=[search_tool, database_tool],
    middleware=[
        ModelRetryMiddleware(
            max_retries=3,
            backoff_factor=2.0,
            initial_delay=1.0,
        ),
    ],
)

from langchain.agents import create_agent
from langchain.agents.middleware import LLMToolEmulator

agent = create_agent(
    model="gpt-5.5",
    tools=[get_weather, search_database, send_email],
    middleware=[
        LLMToolEmulator(),  # 模拟所有工具
    ],
)

from langchain.agents import create_agent
from langchain.agents.middleware import ContextEditingMiddleware, ClearToolUsesEdit

agent = create_agent(
    model="gpt-5.5",
    tools=[],
    middleware=[
        ContextEditingMiddleware(
            edits=[
                ClearToolUsesEdit(
                    trigger=100000,
                    keep=3,
                ),
            ],
        ),
    ],
)

from langchain.agents import create_agent
from langchain.agents.middleware import ProviderToolSearchMiddleware

agent = create_agent(
    model="anthropic:claude-opus-4-8",
    tools=[get_weather, lookup_order],
    middleware=[
        ProviderToolSearchMiddleware(searchable_tools=["lookup_order"]),
    ],
)

from langchain.agents import create_agent
from langchain.agents.middleware import (
    ShellToolMiddleware,
    HostExecutionPolicy,
)

agent = create_agent(
    model="gpt-5.5",
    tools=[search_tool],
    middleware=[
        ShellToolMiddleware(
            workspace_root="/workspace",
            execution_policy=HostExecutionPolicy(),
        ),
    ],
)

from langchain.agents import create_agent
from langchain.agents.middleware import FilesystemFileSearchMiddleware

agent = create_agent(
    model="gpt-5.5",
    tools=[],
    middleware=[
        FilesystemFileSearchMiddleware(
            root_path="/workspace",
            use_ripgrep=True,
        ),
    ],
)

from langchain.agents import create_agent
from deepagents.middleware.filesystem import FilesystemMiddleware

# FilesystemMiddleware 在 create_deep_agent 中默认包含
# 如果你正在构建自定义代理，可以对其进行自定义
agent = create_agent(
    model="claude-sonnet-4-6",
    middleware=[
        FilesystemMiddleware(
            backend=None,  # 可选：自定义后端（默认为 StateBackend）
            system_prompt="当……时写入文件系统",  # 可选的自定义系统提示附加内容
            custom_tool_descriptions={
                "ls": "当……时使用 ls 工具",
                "read_file": "使用 read_file 工具来……"
            },  # 可选：文件系统工具的自定义描述
            tools=["read_file", "ls", "glob", "grep"],  # 可选：允许列表，限制公开哪些文件系统工具
        ),
    ],
)

from langchain.agents import create_agent
from deepagents.middleware import FilesystemMiddleware
from deepagents.backends import CompositeBackend, StateBackend, StoreBackend
from langgraph.store.memory import InMemoryStore

store = InMemoryStore()

agent = create_agent(
    model="claude-sonnet-4-6",
    store=store,
    middleware=[
        FilesystemMiddleware(
            backend=CompositeBackend(
                default=StateBackend(),
                routes={"/memories/": StoreBackend()}
            ),
            custom_tool_descriptions={
                "ls": "当……时使用 ls 工具",
                "read_file": "使用 read_file 工具来……"
            }  # 可选：文件系统工具的自定义描述
        ),
    ],
)

from langchain.tools import tool
from langchain.agents import create_agent
from deepagents.middleware.subagents import SubAgentMiddleware


@tool
def get_weather(city: str) -> str:
    """获取某个城市的天气。"""
    return f"{city} 的天气是晴朗的。"

agent = create_agent(
    model="claude-sonnet-4-6",
    middleware=[
        SubAgentMiddleware(
            default_model="claude-sonnet-4-6",
            default_tools=[],
            subagents=[
                {
                    "name": "weather",
                    "description": "此子代理可以获取城市的天气。",
                    "system_prompt": "使用 get_weather 工具获取城市的天气。",
                    "tools": [get_weather],
                    "model": "gpt-5.5",
                    "middleware": [],
                }
            ],
        )
    ],
)

from langchain.agents import create_agent
from deepagents.middleware.subagents import SubAgentMiddleware
from deepagents import CompiledSubAgent
from langgraph.graph import StateGraph

# 创建一个自定义的 LangGraph 图
def create_weather_graph():
    workflow = StateGraph(...)
    # 构建你的自定义图
    return workflow.compile()

weather_graph = create_weather_graph()

# 将其包装在 CompiledSubAgent 中
weather_subagent = CompiledSubAgent(
    name="weather",
    description="此子代理可以查询城市天气。",
    runnable=weather_graph
)

agent = create_agent(
    model="claude-sonnet-4-6",
    middleware=[
        SubAgentMiddleware(
            default_model="claude-sonnet-4-6",
            default_tools=[],
            subagents=[weather_subagent],
        )
    ],
)

---

<div className="source-links">
<Callout icon="terminal-2">
    [Connect these docs](/use-these-docs) to Claude, VSCode, and more via MCP for real-time answers.
</Callout>
<Callout icon="edit">
    [Edit this page on GitHub](https://github.com/langchain-ai/docs/edit/main/src/oss/langchain/middleware/built-in.mdx) or [file an issue](https://github.com/langchain-ai/docs/issues/new/choose).
</Callout>
</div>