Vibe Coding Architect

Core Problem: The Pitfall of Structural Chaos

A common mistake in Vibe Coding is directly asking an AI to 'build a complete XXX' in one go, hoping for well-structured code. This "open-ended generation" approach almost invariably leads to spaghetti code, tight coupling, haphazard naming, and poor maintainability. Once a project exceeds the prototype stage, this architectural debt rapidly accumulates, making feature iteration cumbersome, debugging costs skyrocket, and ultimately nullifying the benefits of AI-assisted development.

How This Skill Works: An Enforced Structured Workflow

This skill is not a code generator but an enforced engineering methodology. It mandates a clear, ordered dialogue flow to ensure critical design decisions are made before the AI writes a single line of code.

• Trigger & Triage: First, it assesses the task. Simple bug fixes or UI tweaks bypass the skill entirely; adding small features to existing projects jumps to task breakdown; only truly new projects follow the full flow.
• Requirement Clarification (Phase 1): It enforces answers to 6 core questions (e.g., Product Form: Web/CLI/Agent, Project Scale: Demo/MVP, Extensibility Expectations). This acts as a rapid design review, outputting a confirmation table that eliminates ambiguity.
• Architecture Selection (Phase 2): Based on the clarified requirements, it presents 2-3 concrete architectural options (e.g., Monolithic, MVC, Microservices) in a comparative table, evaluating them across dimensions like development speed, maintainability, and scalability. It then provides a clear recommendation with rationale, turning subjective choice into objective trade-off analysis.
• Task Breakdown & Constraints (Phase 3 onward): Only after architecture selection does it proceed to task breakdown. The final AI instruction is built using a 10-item Architecture Prompt checklist, which enforces the inclusion of restrictions (what the AI must not do) and non-functional requirements, effectively 'locking' the architectural decisions into the AI's generation boundaries.

Applicability & Important Considerations

This is a 'heavyweight' workflow designed for 'greenfield projects' and 'major feature additions'. Its value lies in preventing chaos, not in speeding up simple coding tasks.

• Ideal Use Cases: Kicking off a new project, undertaking a major refactor, or introducing a new architectural pattern.
• Not For: Fixing a bug, tweaking a CSS style, or adding a small feature to a project with an already mature and established architecture (where you should follow existing conventions and start directly with task breakdown).
• Key Principle: Once an architecture and tech stack are locked in through this process, they should not be changed lightly during the development cycle. The final output is a detailed architectural prompt with explicit constraints, not immediately executable code. The core philosophy is 'think it through, articulate it clearly, then write the code'.