Abstract
Phishing detection remains a critical challenge due to the evolving nature of social engineering tactics. This study leverages “Tropes” software to extract semantic features from phishing messages, utilizing linguistic and contextual analysis to enhance classification accuracy. By identifying recurring patterns and deceptive language structures, our approach improves phishing detection beyond traditional keyword-based methods. This research highlights the effectiveness of linguistic-based phishing classification and underscores the potential of semantic feature extraction in strengthening cybersecurity defenses against increasingly sophisticated phishing attacks.