From Principles to Practice: A Novel Protocol for Pedagogical Prompt Engineering with Large Language Models

Large Language Models (LLMs) are being used more frequently to provide formative feedback in computing education. However, many of the prompts used to guide these models are based on intuition and lack a structured pedagogical framework. This paper presents the Pedagogical Prompt Engineering Protocol (PPEP), a systematic and theory-informed method for developing effective educational prompts. The protocol follows five structured stages: (1) defining the educational purpose, (2) extracting and integrating foundational pedagogical evidence, (3) translating evidence into design principles, (4) synthesising these principles into core prompt components, and (5) producing a structured prompt design framework. The protocol was implemented in the context of formative feedback for introductory programming. Six computer science experts applied and evaluated the resulting framework, reporting clear improvements in feedback quality, pedagogical coherence, and design transparency. The findings position PPEP as a replicable and theory-informed approach for designing pedagogically aligned prompts for LLM-based applications in computing education.