AI-Driven Autonomous Vulnerability Detection and Remediation Method for Network-Disconnected Cloud Environments

Under military requirements, cloud platforms are often deployed in network-free environments, rendering traditional vulnerability detection and remediation methods difficult to apply. Particularly in the absence of real-time network support, manual repairs exhibit low efficiency and high error rates. To address this challenge, this study proposes a vulnerability detection and repair method based on a locally deployed large-scale deep learning model. Leveraging the model’s zero-shot learning capability and meticulously designed prompt templates, the approach achieves automated vulnerability identification and repair generation in offline environments. Furthermore, the local deployment of the model ensures data security and privacy protection, aligning with military security standards. Experimental results demonstrate that the proposed method effectively enhances the automation level and reliability of vulnerability remediation in network-disconnected scenarios.