Multi Agent

Agent planning (ReAct, Reflexion); tool categories (knowledge, capability, write actions); function calling; multi-agent systems; planning/tool/efficiency failure modes; safety and compound errors; LLM planning debate