Hey Product Hunt! 👋 I'm franklinxkk, a ToB AI product manager based in China. Why I built this: AI coding agents (Claude Code, Codex, Cursor) write code insanely fast. But product delivery is still chaos: - PRDs missing acceptance criteria - Prototypes that can't be tested - Dev handoff that's ambiguous - AI features without runtime governance After 7+ years shipping ToB SaaS and seeing the same problems across CRM, traffic safety, education IT, and healthcare IT projects, I realized: the bottleneck isn't coding speed — it's specification quality. What makes AI Delivery Spec different: | Problem | AI Delivery Spec | |---------|-----------------| | Every feature gets the same template | L0–L3 tier system adapts to complexity | | Prototype ≠ testable | Built-in prototype testability rules | | AI features lack guardrails | AI runtime contract, human gates, rollback | | Coding agents guess business rules | DDD handoff + ac_structured YAML contracts | | Generic skills don't know your domain | 4 replaceable domain modules (CRM, traffic safety, education, healthcare) | Quick start (under 5 minutes): git clone https://github.com/franklinxkk/a... Then tell your AI tool: "Use AI Delivery Spec. Mode=Lite, Tier=L1. Write a light PRD for [your feature]." Works with: Claude Code, Cursor, Codex, Copilot, OpenClaw, Gemini — any AI that reads Markdown. Complementary with GitHub Spec-Kit: spec-kit handles spec→code, AI Delivery Spec handles requirement→spec+prototype. I'd love feedback from: - Product managers shipping ToB/ToG products - Tech leads managing multi-team delivery - Teams adopting coding agents in production Ask me anything! 🙌 🔗 GitHub: https://github.com/franklinxkk/a... 📄 License: Apache 2.0 ⭐ v4.6.2

AI Delivery Spec

The product-side SDD framework for AI-native ToB teams

The product-side SDD framework for AI-native ToB teams