Product Design System is an AI-assisted product design workflow framework designed for product managers. It aims to streamline product design tasks by enabling AI agents to automatically adhere to predefined team specifications, documentation standards, and interaction patterns. This system allows users to fork and customize a general product design framework, benefiting product managers, AI enthusiasts, and design system builders by enhancing collaboration quality and consistency.

How It Works

The system operates on four core mechanisms. First, "Specification as Context" uses AGENTS.md as an AI "work manual" to inject rules and workflows into AI conversations. Second, a "Three-Piece Workflow" generates an HTML prototype, a requirements specification document, and a change log for each new requirement. Third, a self-evolving "Knowledge Base" (product-context/) captures refined design principles and interaction patterns, enriched by AI suggestions and human confirmation. Finally, a "Two-Layer Architecture" separates general, reusable specifications from product-specific details, simplifying customization for new projects.

Quick Start & Requirements

To begin, fork and clone the repository. Configure product-specific rules within AGENTS.md (e.g., product name, platform, component library) and customize the knowledge base files in product-context/ (e.g., business terms, prototype specifications). Load the content of AGENTS.md into your AI tool (Cursor, Claude Projects, ChatGPT) as custom instructions. Initiate collaboration by prompting the AI to create the "three-piece" files for a new requirement. Basic requirements include Markdown proficiency and access to AI tools.

Highlighted Details

• AI-driven product design workflow framework.
• Utilizes structured Markdown for AI context, ensuring adherence to design specifications.
• "Three-piece" output: HTML prototype, requirements document, and change log.
• Knowledge base automatically accumulates design principles and interaction patterns.
• Forkable architecture with general and product-specific layers.

Maintenance & Community

This project is developed and maintained by "懂点儿AI" (Dong Dian Er AI). Feedback and suggestions are welcomed via their WeChat public account. A detailed creation process is available at https://mp.weixin.qq.com/s/TtDijDNzMDXCkLuEufVaww.

Licensing & Compatibility

Released under the MIT License, which is permissive and suitable for commercial use and integration into closed-source projects.

Limitations & Caveats

The effectiveness of this system is contingent on the quality and completeness of the input Markdown context provided by the user. Customizing prototype implementation details may require technical expertise depending on the chosen stack. The knowledge base evolution process necessitates human oversight and confirmation for suggested entries.