Reproducing scores on the posttrainbench website and making submissions

[wise-east]

Can you share the model weights of the agent-trained models on huggingface for reproducibility of the results on the leaderboard?
Is there an official way to submit a model to be recognized on the posttrainbench leaderboard?