Observation
Domain: Agent Fundamentals
🎯 Motivation
An agent perceives its environment through observations - tool outputs, user messages, system state. Observations inform decisions.
📋 Concrete Example
The agent calls get_weather(HK) and observes: ‘28C, sunny.’ This observation updates its understanding. Next action depends on this observation.
🔗 Analogy
A driver checking mirrors and road signs - each check is an observation. The driver adjusts behaviour based on what they observe. Agents observe tool outputs to adjust their actions.