A Field Report on AI-Assisted Development of an Electron+SQLite Hospital Management System: A Taxonomy of 44 Pitfalls and 20 Lessons Learned

This experience report documents the development of a desktop hospital management system using Electron and SQLite over a period of six versions (V1.0 to V2.0). The development process was conducted through a series of AI-assisted sessions, where a human developer and an AI assistant collaborated iteratively. Through this process, we encountered and documented 44 distinct pitfalls that disrupted development. This report contributes a taxonomy of these pitfalls, organized into eight categories: PowerShell and Terminal Issues, Electron Framework Problems, SQLite Database Challenges, Build and Packaging Failures, Security Vulnerabilities, Frontend and SPA Errors, JavaScript Syntax Issues, and Financial System Design Flaws. For each category, we present the symptoms, root causes, and the golden rules that were established to prevent recurrence. We also derive 20 golden rules for practitioners. Our findings suggest that while AI assistants can significantly accelerate software development, the collaboration process itself introduces a new class of pitfalls that must be managed through disciplined documentation and context management.