I built a static analysis tool that audits LangChain agents for excess tool permissions before they ship. Here's what it found on typical agent patterns developers use every day.
When you write this:
agent = initialize_agent(
tools=[GitHubTool, SlackTool, SQLDatabaseTool],
llm=llm,
agent_kwargs={"system_message": "You summarize pull requests."}
A static AST scan of three open-source TypeScript agent codebases found 669 tool calls with real side effects. 553 had no guard…
If you've built anything with LangGraph past the demo stage, you've probably had the same uneasy...
When I started building Agent_Sudo, I thought the hard part would be approvals and policy...
LangChain agents are built on LangGraph: the model calls tools in a loop until it returns a final...
Look, LangChain was great to get the ecosystem started 0 to 1. But let's be honest: taking LangChain...
See Inside Your LangGraph Agents Open-source LangChain/LangGraph tracing — drop in two...
