Your question is State Management for Long Running Agents. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
You're building a long-running agentic workflow that spans many turns, tool calls, and intermediate decisions. Over time, the conversation history, tool outputs, and working memory grow beyond the model's context window. You need a way to preserve the right state, drop the wrong state, and keep the agent reliable as it continues operating.
How would you handle context window limitations and state management in a long-running agentic workflow?