Mechanism design theory, introduced by 2007 Nobel laureates Hurwicz, Maskin, and Myerson, has guided economic institutions worldwide by efficiently allocating resources to achieve desirable outcomes. However, current mechanism design theories rely on trusted third parties. In reality, we regret seeing that mutually beneficial transactions fail to occur due to trust issues, and social welfare cannot be fully optimized. Blockchain technology, known as the trust machine in cyberspace, can replace trusted third parties with tamper-proof algorithms. However, how blockchain will enable mechanism design remains to be studied. More importantly, the past decades have witnessed an emerging digital economy with an increasingly complex system. New application scenarios have inspired new questions. How can we analyze big data in a computationally efficient way to draw causal inferences for policy advice? How can we extend the foundation of mechanism design theory to create trust in cyberspace? How can we design experiments to identify more realistic behavioral assumptions underpinning our theory? To answer these important but challenging questions, this research agenda explores trust mechanism design on blockchains through an interdisciplinary study. First, we apply the method of machine learning and causal inference to evaluate the efficiency and fairness of existing mechanisms by analyzing the historical data of natural experiments on the blockchain. Then, we investigate the design principles of trust mechanisms on blockchains by an integrated method of algorithmic game theory, mechanism design, and reinforcement learning. Finally, we research the behavioral foundations of trust mechanism design on blockchains using behavioral experiments, especially the methods in the human-computer interaction and bounded rationality literature. This research agenda has the potential to expand the theoretical foundation of trust mechanism design and empower the digital transformation of Industry 4.0.