📄 中文摘要
在使用AI编码助手为应用层添加新用例时，生成的代码迅速完成，但同时违反了三个架构规则。这些规则在之前的讨论中详细阐述，强调了清洁架构的重要性。为了应对LLM生成代码带来的挑战，作者提出了使用自定义Detekt规则、专门的AI代理和规范驱动开发的方法，以确保清洁架构的完整性，即使在LLM编写代码的情况下也能保持架构的稳定性。
LLM生成的代码破坏了清洁架构：我如何通过Detekt在构建时强制执行它
📄 English Summary
LLM-Generated Code Keeps Breaking Clean Architecture: How I Enforce It at Build Time with Detekt
An AI coding assistant was tasked with adding a new use case to the application layer, completing the code in under a minute. However, this generated code broke three architecture rules that had been previously established in detail. To address the challenges posed by LLM-generated code, the author proposes the use of custom Detekt rules, specialized AI agents, and specification-driven development. These strategies aim to maintain the integrity of Clean Architecture, ensuring that architectural standards are upheld even when the code is written by an LLM.