Define how the agent should behave and respond to users.

Choose the LLM model this agent will use. If not selected, a default will be used.

Control how the agent manages its memory.

Maximum number of memory items before summarization (default: 1000).