Conformal Off-Policy Prediction for Multi-Agent Systems