If an agent task takes 2 minutes to complete, how do you architect the API so the user's browser connection doesn't time out?

If a tool returns a massive 50MB JSON file, you can't feed that to the LLM. How do you "summarize" or "filter" tool observations for the agent?

How do you implement "Long-Term Memory" for an agent so it remembers a user's preference from a conversation that happened three weeks ago?

A user asks "What were the sales in 2023?" How do you prompt the LLM to separate the semantic search ("sales") from the metadata filter ("year == 2023")?