gszhangwei/open-spdd

What it solves

OpenSPDD addresses the problem of "disposable" AI coding prompts and vague plan documents that lead to AI improvisation, lack of traceability, and design drift. It transforms AI prompts into executable design contracts, ensuring that the AI follows precise architectural patterns and coding standards rather than just generating a high-level task list.

How it works

The project provides a cross-platform CLI tool that integrates with various AI coding environments (like Cursor, Claude Code, and GitHub Copilot). It implements the REASONS Canvas, a 7-dimensional structured design framework (Requirements, Entities, Approach, Structure, Operations, Norms, and Safeguards) to define the "why," "what," and "how" of a feature.

Users follow a structured workflow: strategic analysis $\rightarrow$ REASONS Canvas generation $\rightarrow$ code generation $\rightarrow$ reverse synchronization. The CLI tool manages the templates and command files required to trigger these steps within the AI tool's native command system.

Who it’s for

• Enterprise developers needing design-implementation traceability and long-term maintainability.
• Development teams seeking unified AI coding standards to reduce style conflicts.
• Architects performing complex refactoring where strict operation orders are necessary to prevent dependency chaos.

Highlights

• Bidirectional Sync: Allows users to sync code changes back into the design documents to keep the prompt and implementation aligned.
• Cross-Platform Support: Compatible with Cursor, Claude Code, GitHub Copilot, Antigravity, OpenCode, and Codex.
• Structured Framework: Uses the REASONS Canvas to specify exact method signatures, error handling, and DI patterns instead of vague instructions.
• Auto-detection: Automatically identifies the active AI coding environment to configure the correct command directories.