A catalogue of reusable patterns for building robust AI agents: ReAct, Reflection, Tool Use, Plan-and-Execute, and more.
Five passes over the same idea, each from a different angle. Do them in order, or jump to whichever you need.
Just as Gang-of-Four patterns gave us a shared vocabulary for OOP design problems, agent design patterns give us reusable solutions for LLM orchestration. ReAct interleaves reasoning with action. Reflection adds self-critique. Plan-and-Execute separates planning from execution. Understanding these patterns helps you avoid reinventing solutions to solved problems.
