pmai-proposal

Generates versioned Product Proposal documents with evidence discipline and atomic PRODUCT.md synchronization.

4|Updated May 26, 2026
One-click install
npx skills add https://github.com/YYG501/PM-AI-Harness --skill pmai-proposal-yyg501
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: pmai-proposal
Source: https://github.com/YYG501/PM-AI-Harness/tree/main/skills/proposal
Command: npx skills add https://github.com/YYG501/PM-AI-Harness --skill pmai-proposal-yyg501

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Product managers often start building before clarifying why a product should exist, who it serves, and what the MVP must prove. This Skill turns scattered product thinking into a complete, reviewable Product Proposal with explicit evidence levels, falsifiable judgments, and a versioned decision trail. ## Core Features & Use Cases - Complete Proposal Generation: Produces a full Product Proposal covering users, problems, product mechanism, boundaries, value chain, MVP case rundown, and downstream handoff summary. - Version & Supersede Management: Creates new versions (v1, v2...) for direction changes, maintains docs/proposals/INDEX.md, and atomically syncs the stable baseline into PRODUCT.md with a machine contract. - Evidence Discipline: Separates confirmed facts, reasoned inferences, and unverified assumptions, and requires a most-dangerous premise with a falsification path. - Use Case: A solo PM starting a new AI product runs the proposal flow to decide whether the product is worth building, defines the MVP validation gate, and hands the first buildable result to the design stage. ## Quick Start Ask the assistant to run /pmai-proposal to clarify the product direction and produce a complete, reviewable Product Proposal for your project.

Frequently Asked Questions about pmai-proposal

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I write a product proposal before starting development?

Run the proposal workflow to produce a complete Product Proposal covering users, problems, product mechanism, boundaries, value chain, and an MVP case rundown. It separates confirmed facts from assumptions and ends with a downstream handoff summary for the design stage.

What should a Product Proposal include for an AI product?

An AI product proposal must justify AI necessity per specific task, define human-AI division of labor, and state how AI errors are detected and recovered. Deterministic work like permissions, calculations, and audit trails stays with deterministic software.

When should I create a new proposal version instead of editing?

Create a full new version whenever positioning, target users, value, boundaries, or MVP change. Finalized versions are never rewritten in place; the old version keeps its content with a superseded status header and the index tracks one current version.

Can I skip the proposal step for an existing project?

Yes, but only when the intake flow has verified complete materials, PRODUCT.md records the in-repo basis with PM confirmation date, and the machine state is equivalent_baseline. Otherwise entering the skill requires completing a full proposal.

What is the difference between a proposal and a PRD in this workflow?

The proposal answers why the product is worth building and what the MVP must prove, while PRDs and functional specs are handled by the spec-writing stage. Module objects, fields, states, and page interactions belong to the design stage, not the proposal.