Ensuring the Maintainability and Supportability of "Vibe-Coded" Software Systems: A Framework for Bridging Intuition and Engineering Rigor

This paper addresses the emerging concept of "vibe-coding"—the translation of intuitive feelings or high-level intentions directly into software code. While potentially accelerating development or enabling novel forms of creation, such code inherently risks being opaque, poorly understood, and difficult to maintain. We propose the Intent-Driven Explicable Architecture (IDEA) framework, a novel approach designed to ensure that vibe-coded software remains maintainable, understandable, and supportable throughout its lifecycle. IDEA integrates techniques for formalizing intuitive inputs, constraining code generation using established software engineering principles, automatically generating explanations linking code to intent, and incorporating rigorous human-in-the-loop validation. We argue that by structuring the translation process and embedding traceability and explicability, IDEA mitigates the primary risks associated with intuition-driven code generation, paving the way for its responsible exploration.