This study examines the integration of prompt engineering and fine-tuning strategies with metacognitive skills to enhance tertiary-level written assessments in the Generative AI (GenAI) era. Mixed-methods approaches were employed, including a Fuzzy Delphi technique (n = 30), an online survey (n = 171), and a focus group discussion (n = 14) involving academicians from seven different fields of study across Malaysia, Indonesia, Thailand, Singapore, and the Philippines. Findings indicate that fine-tuning based on metacognitive theory improves shaping behavior, interaction strategies, and natural language querying abilities. This prepares students to adapt and benefit from personalized feedback that supports deeper learning in tertiary-level written assessments. The study offers practical guidelines for academicians to apply metacognitive fine-tuning strategies in prompts. Overall, this contributes to the advancement of written assessment design by elucidating the critical importance of metacognitive integration in enhancing both tuning and fine-tuning of prompting skills in the presence of GenAI tools.