Meet the newest @LangChain trace engineer in Chicago on June 22nd and see @pollenrobotics reachy pull up traces, explain them, and do all the cool things a deep agent can do!
Oh and @hwchase17 and I will be talking on deep agents, AI in the Enterprise, and a ton more too...
When Box transitioned Box AI into a multi-step agentic system, we chose @LangChain's LangGraph over building a custom execution engine from scratch.
Our full architectural breakdown covers how agents are defined and compiled at runtime, how the Global Agent spawns sub-agents
.@Box Agent is built on Deep Agents.
✅ Cross-library search
✅ Multi-doc synthesis
✅ Structured reports
✅ All within Box's existing security and permissions model
everybody's talking about loops!! how can you instrument them with langchain?
1. token loop supported by a model (choose any model with langchain)
2. create_agent gives you the agent loop (model + tools + repeat until done)
3. deepagents gives you the self verification loop
4 ways you can create skills in LangSmith Fleet:
✅ Create with AI: Open Chat and describe what you want the skill to do.
✅ Generate during agent creation: When you create a new agent, Fleet automatically generates relevant skills if the agent would benefit from them.
✅ Start
Let’s talk about the agent lethal trifecta, coined by @simonw
1️⃣Access to sensitive data
2️⃣Exposure to untrusted content
3️⃣The ability to communicate externally
If all 3 apply to your agent, it needs a sandbox.
langchain.com/blog/how-to-ch…?
we revamped our skills docs this week!
what are they missing? what questions do you have about skills that are unanswered here?
docs.langchain.com/oss/python/dee…