Langchain Middleware

For production LangChain agents that need human oversight on risky operations like sending emails or deleting data. Sets up human-in-the-loop middleware that pauses execution before dangerous tool calls, letting humans approve, edit arguments, or reject with feedback. Also covers custom middleware hooks for logging, retries, and error handling. The resume pattern with Command objects handles the back-and-forth flow cleanly. Requires a checkpointer and thread_id configuration, which trips up developers initially but makes sense once you need to pause and resume agent execution across different processes.

skills.sh installs: 17K
GitHub stars: 1.3K
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add langchain-ai/langchain-skills --skill langchain-middleware --agent claude-code

Installs into .claude/skills of the current project.

Files

Middleware patterns for production LangChain agents:

HumanInTheLoopMiddleware / humanInTheLoopMiddleware: Pause before dangerous tool calls for human approval
Custom middleware: Intercept tool calls for error handling, logging, retry logic
Command resume: Continue execution after human decisions (approve, edit, reject)

Requirements: Checkpointer + thread_id config for all HITL workflows.

Human-in-the-Loop

Set up an agent with HITL middleware that pauses before sending emails for approval.

from langchain.agents import create_agent
from langchain.agents.middleware import HumanInTheLoopMiddleware
from langgraph.checkpoint.memory import MemorySaver
from langchain.tools import tool

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

SKILL.md on GitHub