A Practitioner's Guide to Multi-turn Agentic Reinforcement Learning