A breakdown of indirect prompt injection attacks using real-world cases (emails, code comments, diagrams). Introduces the CFS model (Context, Format, Salience) to explain what makes some payloads more likely to succeed. Comments
You must log in or register to comment.