Decentralized finance (DeFi) protocols, especially lending and credit delegation markets, carry complex systemic risks that traditional models often fail to capture. In this talk, I will present how agent-based modeling (ABM) offers a powerful lens to simulate and analyze emergent behaviors, liquidity shocks, and governance parameter impacts in these systems. Using concrete case studies, including my work on optimizing protocol parameters and analyzing liquidation dynamics, I will demonstrate how quantitative ABM can expose hidden vulnerabilities, guide parameter design, and improve the resilience of DeFi markets. This session will be highly practical, focused on actionable insights for researchers, protocol designers, and risk managers.