Reflection
Domain: Agent Fundamentals
🎯 Motivation
After completing a task, the agent reflects on what went well, what failed, and what to improve. This metacognition improves future performance.
📋 Concrete Example
After a failed coding attempt: ‘The error was a type mismatch. I should have checked types first. Next time, I will add type checking before running.’ The reflection improves future attempts.
🔗 Analogy
A player reviewing game footage - they watch their mistakes, identify what to improve, and play better next time. Reflection is the agent reviewing its own ‘game footage.‘