Build your PMF context layer — ICP, value prop, MVP PRD — then validate with landing pages, conversations, or BMAD.
npx claudepluginhub adidacta/pmf-detectiveBuild your PMF context layer - reference files that capture the 'WHY' behind your product. Claude uses these files when building anything for your product.
Share bugs, ideas, or general feedback.
A Claude Code plugin that walks you from idea to validation. Define your ICP, craft your value proposition, scope your MVP, then validate — with landing pages, Mom Test conversations, or the BMAD Method.
Created by Adi Shmorak, The P/MF Detective
Most builders skip straight to code. Plan PMF forces you to answer the hard questions first — who is this for, why would they care, and what's the smallest thing you can build to prove it? The output is a set of pmf/ files that Claude references when building anything for your product.
/install github:adidacta/pmf-detective
Plan PMF guides you through 4 sections, then routes you to execute:
1. ICP → pmf/icp.md (assumption)
2. Value Proposition → pmf/value-prop.md (assumption)
3. MVP PRD → pmf/mvp.md (assumption)
4. Validation Plan → pmf/validation-plan.md (decides HOW to test)
│
├─→ Landing Page (signups / pre-orders)
├─→ Outreach (Mom Test conversations)
└─→ Build & Test (BMAD Method)
The first 3 sections are assumptions. The validation plan decides how to test them, then routes you to the right execution path.
Tell Claude what you need:
Or use slash commands:
| Command | What it does |
|---|---|
/plan-pmf | Build your full context layer (guided flow) |
/pmf-status | Check context completion status |
/update-icp | Refine your ICP |
/update-value-prop | Refine your value proposition |
/update-mvp | Refine your MVP scope |
/generate-assets landing | Generate a landing page from your context |
Define who your customer is. Generates 3 hypotheses, researches each in parallel, and helps you pick the strongest:
Output: pmf/icp.md
Craft messaging that makes your ICP stop and take action:
Output: pmf/value-prop.md
Scope what to build by working backwards from the core aha moment:
The output is a high-level PRD that any AI coding agent can use as project context.
Output: pmf/mvp.md
Decide how to test your assumptions before going all-in:
Output: pmf/validation-plan.md
After your validation plan, Plan PMF routes you to one of three paths:
Generates a React + Vite landing page from your context files. Uses your Callout, Magnet, Path to Aha, CTA, and ICP language. Includes SEO and social tags.
/generate-assets landing
Creates an outreach plan with Mom Test interview questions. Grounded in your ICP's channels, language, and pain points.
Bridges your PMF context into the BMAD Method for AI-driven development. Your pmf/ files become the product definition — BMAD handles architecture, sprint planning, and implementation.
All outputs live in a pmf/ folder in your project:
pmf/
├── icp.md # Who your customer is
├── value-prop.md # Why they should care (Callout + Magnet)
├── mvp.md # MVP scope — aha moment, features & requirements
├── validation-plan.md # How you'll test + success thresholds
└── outreach-plan.md # (if conversations) Mom Test questions + channels
Claude references these files when building anything for your product — landing pages, copy, features, architecture, documentation.
Should you commit pmf/? Yes. These files are your product context. Committing them means every team member and every Claude session shares the same foundation. Update them as you learn from validation.