The rapid integration of generative artificial intelligence (AI) tools in programming education has transformed how students approach coding tasks. In web development courses, students are becoming increasingly reliant on AI systems to generate, refine, and debug code. This shift raises an important pedagogical question: does effective prompting represent a meaningful cognitive skill, or is it merely the act of entering instructions into an AI system? This study investigates prompting competency as a measurable construct and examines its relationship with student project performance in undergraduate web programming education. The research was conducted in a web development course in which students individually designed and developed an educational hybrid game. A mixed-methods approach was employed, combining analysis of students’ AI prompt logs, project performance assessments, and reflective reports. Prompting competency was operationalized through a Prompting Competency Index (PCI) consisting of five dimensions: 1) prompt specificity, 2) contextual framing, 3) iterative refinement, 4) debugging strategy, and 5) critical evaluation of AI-generated outputs. Quantitative analysis examined the relationship between PCI scores and project performance, while qualitative analysis explored students’ metacognitive strategies during prompt formulation. The findings indicate that higher-performing students demonstrate structured, iterative, and reflective prompting behaviours. These behaviours align with higher-order cognitive processes such as computational thinking, debugging strategies, and self-regulated learning. The results suggest that prompting does not merely constitute mechanical interaction with AI systems, but an emergent form of digital literacy that can be observed and assessed, highlighting the importance of developing students’ awareness of effective prompting strategies in AI-assisted programming tasks. This study contributes to ongoing discussions on AI-assisted learning by positioning prompting competency as a pedagogically relevant capability in contemporary computing education.