Back to directory

OpenSPDD

A methodology and cross-platform CLI that turns AI coding prompts into structured design contracts with explicit constraints and dependencies, kept in sync with the implementation through reverse synchronization.

OpenSPDD

OpenSPDD stands for Structured Prompt-Driven Development. It is a methodology and command-line tool that treats an AI coding prompt not as a one-off instruction but as a design contract: a structured document with explicit constraints on what an agent may and may not do, precise details such as method signatures and error handling, and an explicit execution order for dependent steps.

Its core artifact is the REASONS Canvas, a seven-part structured framework covering design decisions, the implementation path and quality guardrails. A sync command lets the design document and the code it describes be reconciled in either direction as the implementation diverges from the plan, following the principle of fixing the prompt first and then updating the code.

It works with several AI coding tools, including Cursor, Claude Code, GitHub Copilot, Antigravity, OpenCode and Codex, auto-detects which one is in use, ships as a single Go binary with all templates embedded, and installs with Homebrew or Go install.

More in Agent skills

See all 581
  • Provenance gate

    github.com

    A deterministic authorization gateway that sits between an AI agent and its tools, tracing where each parameter of a side-effect call came from, and blocking the call or asking a human when the source does not check out.

  • A markdown framework of system rules that restates Isaac Asimov's Four Laws of Robotics as a set of prioritized loss calculations for large-language-model agents to follow, published as a discovery file for agents and crawlers to read.

  • Vitruvius

    github.com

    An engineering research agent that runs a discover, read, synthesize, verify and review loop over technical questions, reading sources directly rather than inferring from titles, and publishing its own scored verifier benchmark, weaknesses included.