saadrasool43/NEC-4-Contract-Risk-Detection-using-LLMs-via-Open-AI-API: Evaluating Prompt Engineering Strategies for Large Language Model–Based Risk Detection in Construction Contracts.

This release provides the full replication package for the study: "Evaluating Prompt Engineering Strategies for Large Language Model–Based Risk Detection in Construction Contracts." The repository contains: The Python scripts used to generate LLM outputs via the OpenAI API The finalized prompt templates for all five prompting strategies (baseline, direct, role-based, guided, chain-of-thought) The complete set of 30 LLM-generated risk analysis outputs (6 synthetic NEC4-style Z-clauses × 5 prompt types) Documentation enabling full reproducibility of the experimental procedure All experimental parameters, prompts, and outputs correspond exactly to those reported in the submitted manuscript. This release is archived to support transparency, reproducibility, and peer review in accordance with best practices for computational research in engineering informatics.