With the rising of Artificial Intelligence (AI) and its applications it became crucial for rapid development applications, companies start the adoption of AI assistance and plugin to help developers writing code faster, vibe coding is a new term that refer development of applications using prompts to generate code by conversation, although this help a lot it raise a new concern about the security of such applications, the need of automated vulnerabilities scanning it become essential, DevOps engineers start to implement static code analysis tools into pipelines, the main disadvantage of such tools that it produce large number of false alarms and that killed the whole idea of continuous delivery, to solve the problem of false positive alarms this study suggest a new methodology that employe

Терро М. (науч. рук. Заколдаев Д.А.) Automated security scanning in ci/cd pipelines using large language models (llms) // Сборник тезисов докладов конгресса молодых ученых. Электронное издание. – СПб: Университет ИТМО, [2026]. URL: https://kmu.itmo.ru/digests/article/16334