This research aims to investigate the vulnerabilities of existing federated learning frameworks against backdoor attacks and to develop a novel backdooring approach. The study focuses on advancing a new technology in this area. The research uses textual data, with all data (i.e., source code) being independently developed.