Show HN: Rules that stop AI coding agents from breaking working code (github.com)

A newly shared framework aims to enhance the collaboration between AI coding agents, such as Claude Code, Cursor, and Codex, and human developers by implementing structured rules to prevent these systems from inadvertently breaking functioning code. The framework outlines specific directives to guide AI behavior, emphasizing the importance of constraints over mere prompting, as the complexity of larger codebases can lead to unintended consequences. Key rules include prohibiting modifications to files not explicitly referenced and requiring the completion of tests before changes are deemed finalized. This initiative is significant for the AI/ML community as it addresses a critical limitation when integrating AI into software development workflows. By creating a systematic approach with clear safety measures, the framework not only reduces the risk of introducing new errors but also emphasizes the need for thoughtful planning and verification before code alterations. The proposed rules encourage a more controlled interaction with AI models, promoting a balance between automation and human oversight, which could lead to more reliable outcomes in software development.