Vulnerabilities of Foundation Model Integrated Federated Learning Under Adversarial Threats