While a basic Large Language Model (LLM) agent—one that repeatedly calls external tools—is easy to create, these agents often struggle with long and complex tasks because they lack the ability to plan ...