Desenvolvimento de um bot de captura de mensagens do Telegram para inteligência de ameaças cibernéticas

Abstract

This work proposes the development of a backend system for the collection and analysis of messages from public Telegram channels, aiming to support cyber threat intelligence. To achieve this, a bot was created using the NestJS framework, the TypeScript language, the Prisma ORM, and the PostgreSQL database, with containerization via Docker. The application collects messages both automatically and manually, enabling their storage and subsequent classification regarding their relevance to the cybersecurity context through supervised machine learning techniques. During the experiments, 760 messages were collected, of which approximately 16% were classified as relevant. The results demonstrate the feasibility of the application in assisting the identification of malicious content on social networks, also proposing its future integration with other communication platforms.