Integrating Large Language Models into Cybersecurity: Prompt-Based Code Analysis
2nd International Symposium on AI-Driven Engineering Systems, ISADES 2026, Hybrid, Mbale, Uganda, 19 - 20 Haziran 2026, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Doi Numarası: 10.1109/isades69945.2026.11608208
- Basıldığı Şehir: Hybrid, Mbale
- Basıldığı Ülke: Uganda
- Anahtar Kelimeler: code security analysis, cybersecurity, large language models, prompt engineering, vulnerability detection
- Dokuz Eylül Üniversitesi Adresli: Evet
Özet
This study presents a 'Prompt-Based Security Analyzer' system to automate the software security process using Large Language Models (LLMs). To address the contextawareness limitations of traditional Static Application Security Testing (SAST) tools, Google Gemini series, OpenAI's GPT series, and Anthropic Claude series models are integrated into the system via special prompt engineering techniques. The developed Python-based system not only detects critical vulnerabilities like SQL Injection and Cross-Site Scripting (XSS) in source code, but it also provides developers with a comparative interface for suggesting safe code blocks (fix suggestions). Experimental results have shown that the system provides high accuracy in vulnerability detection and significantly accelerates the secure software development process.