The increasing reliance on digital and automated systems in the maritime industry has introduced significant cybersecurity risks to ship operations. While prior research has examined cyber risk management, systematic modeling of attack scenarios remains limited. This study addresses this gap by employing attack trees to systematically analyze potential cyberattack pathways against modern ships. Drawing insights from the literature and interviews with 10 experts, this research identifies key vulnerabilities in navigation, operational, and communication systems. The study outlines 20 attack scenarios across three primary objectives: crashing, capsizing, and immobilizing a ship. Expert validation emphasizes the real-world feasibility of these threats, with cyberattack scenarios involving crashing and immobilization deemed the most plausible, while capsizing was considered unlikely due to robust ship designs. The findings highlight the importance of strengthening network security, implementing redundancy measures, and enhancing crew training to mitigate cyber risks.in Expert feedback helped validate these insights, underscoring their relevance to real-world maritime operations. This research presents a systematic threat modeling framework that supports effective risk assessment and cybersecurity strategy development, offering a valuable contribution to the advancement of maritime cybersecurity practices.

LLM agents are widely used as agents for customer support, content generation, and code assistance. However, they are vulnerable to prompt injection attacks, where adversarial inputs manipulate the model's behavior. Traditional defenses like input sanitization, guard models, and guardrails are either cumbersome or ineffective. In this paper, we propose a novel, lightweight defense mechanism called Polymorphic Prompt Assembling (PPA), which protects against prompt injection with near-zero overhead. The approach is based on the insight that prompt injection requires guessing and breaking the structure of the system prompt. By dynamically varying the structure of system prompts, PPA prevents attackers from predicting the prompt structure, thereby enhancing security without compromising performance. We conducted experiments to evaluate the effectiveness of PPA against existing attacks and compared it with other defense methods.