What problem does it solve? Building production agents with LangChain requires knowing the current recommended patterns, and outdated approaches lead to broken loops, lost conversation state, and missing human approval controls. This Skill provides the canonical patterns for creating agents with create_agent, defining tools, and adding middleware. ## Core Features & Use Cases - Agent Creation: Build agents with create_agent using model strings or configured model instances, system prompts, and tool lists in Python or TypeScript. - Middleware & Human-in-the-Loop: Add HumanInTheLoopMiddleware for approval workflows, custom wrap_tool_call hooks, and resume interrupts with Command objects. - State & Reliability Fixes: Apply checkpointers with thread_id for memory, recursion_limit for loop control, and correct result message access patterns. - Use Case: A developer building a customer support agent can define Salesforce lookup tools, require human approval before issuing refunds, and persist conversation state across sessions. ## Quick Start Create a LangChain agent with create_agent that uses a weather lookup tool and remembers the conversation across invocations.
