ABSTRACT Prompting is becoming a foundational skill in the age of generative AI, yet little is known about what makes someone a good prompter. This study introduces a multi‐dimensional framework of prompt literacy and investigates it through an eye‐tracking experiment. Sixty participants were invited to create short stories using the Doubao platform, while their gaze behavior was recorded. By integrating cognitive data, experiential reports, and sociocultural backgrounds, we analyze how prompt construction patterns relate to satisfaction, perceived difficulty, and creative output. Our findings provide insight into the competencies that define good prompting and offer practical implications for prompt training and interface design.