Multi-Agent Intention Progression with Reward Machines