一句话：提出结构化规约驱动工程（SSDE）范式，以结构化规约替代自然语言提示引导 LLM 生成仓库级代码，并通过试点研究验证其可行性与可验证性优势。
当前最先进的大语言模型在函数级代码生成上表现优异，但扩展到仓库级系统时输出质量显著下降；仅依赖自然语言提示的工作流存在固有歧义、缺乏可验证性。为此，作者提出结构化规约驱动工程（SSDE）：以结构化规约等结构化产物作为 LLM 输入来指导仓库级代码生成，认为这既能提升生成质量，又能带来更好的可验证性。为检验该设想的可行性，作者开展了一项试点研究，使用五种 LLM 为三个软件系统生成 Model-View-Controller（MVC）业务逻辑，并据此讨论了 SSDE 的潜力、挑战与未来路线图。该工作属于立场与愿景型研究，尚未给出大规模量化评估。
State-of-the-art Large Language Models (LLMs) excel in code generation at the function level. However, the output quality significantly declines when scaling to repository-level systems. Current workflows relying only on natural language prompts suffer from inherent ambiguity and a lack of verifiability. To address this, we propose structured spec-driven engineering (SSDE), a paradigm that leverages structured artifacts to guide LLM generation. We argue that structured specifications as LLM inputs make high-quality, repository-level code generation a tangible goal, while at the same time offering superior verifiability, leading to significant potential for improvement. We first investigate the feasibility of this vision through a pilot study generating Model-View-Controller (MVC) business logic for three software systems using five LLMs, and then highlight the potential, challenges, and future roadmap for SSDE.
DOI 原文 · arXiv · PDF（开放获取） · DBLP
@inproceedings{abs-2605-02455,
title = {LLM-Assisted Repository-Level Generation with Structured Spec-Driven Engineering},
author = {Shuzhao Feng and Boqi Chen and Brett H Meyer and Gunter Mussbacher},
booktitle = {FSE},
year = {2026},
doi = {10.1145/3803437.3805567}
}