Supply Chain refers to vulnerabilities in the development and deployment processes of Large Language Models (LLMs), where compromised third-party components—such as pre-trained models, datasets, or plugins—can introduce security risks like backdoors, biases, or system failures, potentially leading to unauthorized access or malicious behavior.