This workshop focused on the security risks associated with Large Language Models (LLMs), particularly in chatbot and RAG-based systems. Participants explored real-world attack scenarios such as prompt injection and data leakage through live demonstrations and guided exercises. The session highlighted why common defenses fail and introduced effective mitigation strategies, including prompt injection detection, guardrails, and secure system design. Participants left with a strong security-first mindset and practical techniques to safeguard LLM applications.