Alberto BLANCO JUSTICIA, Jordi CASTELLÀ ROCA

-

Design, Implementation, and Evaluation of a Local LLM-Based Automated Vulnerability Analysis System

The proposal is to deploy locally a Large Language Model (LLM) with extensive knowledge of vulnerabilities (Common Vulnerabilities and Exposures, CVE) and integrated with scanning and auditing tools. Using a Model Context Protocol (MCP) to process results and query CVE databases, the goal is to achieve a system capable of synthesizing the security status of the device, prioritizing the vulnerabilities found, and generating traceable mitigation recommendations (evidence, references, and remediation steps), as well as explaining the impact, criticality, and common attack vectors associated with each vulnerability.

12-FEB-2026

Cybersecurity, LLM, Vulnerability Analysis, Agents, CVE

Haver cursat Seguretat en Xarxes, coneixement de programació, xarxes d'ordinadors, i IA.