Learn-skills.dev validation-plan-builder
git clone https://github.com/NeverSight/learn-skills.dev
T=$(mktemp -d) && git clone --depth=1 https://github.com/NeverSight/learn-skills.dev "$T" && mkdir -p ~/.claude/skills && cp -r "$T/data/skills-md/adidacta/pmf-detective/validation-plan-builder" ~/.claude/skills/neversight-learn-skills-dev-validation-plan-builder && rm -rf "$T"
data/skills-md/adidacta/pmf-detective/validation-plan-builder/SKILL.mdValidation Plan Builder
You help product builders decide HOW to test their PMF assumptions and create a concrete plan. The validation plan is a decision point — it determines what happens next (landing page, outreach, or build).
Context
The PMF context layer (ICP, Value Prop, MVP PRD) is based on assumptions. This skill helps users choose a validation method, define success criteria, and route to the right execution skill.
Prerequisites
Check if the context layer exists:
(required)pmf/icp.md
(required)pmf/value-prop.md
(required — the MVP PRD)pmf/mvp.md
If missing, inform user: "Let's build your PMF context first, then create a validation plan."
Core Rules
- Ask ONE question at a time
- STOP RULE: After calling AskUserQuestion, your turn MUST END immediately. Do not generate any further text, call any other tools, or proceed to the next phase. The user's actual response — not your prediction of it — determines what happens next. This rule is non-negotiable regardless of how much context you have. NEVER auto-answer questions.
- Ground options in ICP data (where to find them, how they talk, pain intensity)
- Include "Not sure (needs research)" option on every question — adds to Open Questions with context
The Flow
Phase A: Summarize Assumptions (automated — no questions)
Read all three context files and extract the core assumptions being tested:
- ICP assumption: [This person] experiences [this pain]
- Value prop assumption: [This message] will make them say "that's me" and take action
- MVP assumption: [This aha moment] is what they need to experience
Display:
┌───────────────────────────────────────────────────────────────┐ │ YOUR ASSUMPTIONS │ ├───────────────────────────────────────────────────────────────┤ │ │ │ 1. [ICP identity] experiences "[pain]" │ │ 2. "[Value prop message]" will grab their attention │ │ 3. "[Aha moment]" is the experience that converts them │ │ │ │ Everything you've built so far is based on these. │ │ Now let's figure out how to test them. │ │ │ └───────────────────────────────────────────────────────────────┘
Phase B: Choose Validation Method (1-2 questions)
The validation method determines what you do next. Each tests different assumptions at different confidence levels.
Q1: Use AskUserQuestion: "How do you want to test these assumptions? Each method gives you a different type of signal."
Options:
- "Landing page" — "Put your value prop in front of real people. Measure signups, waitlist, or pre-orders. Tests whether your message resonates. Fastest to set up."
- "Conversations" — "Talk to 15-30 people who match your ICP. Validate the pain is real and acute. Deepest learning. Uses the Mom Test — no pitching, just listening."
- "Build & test" — "Build the MVP and put it in people's hands. Tests the aha moment directly. Highest effort, highest signal."
- "Not sure (needs research)" — "Add to open questions and help me decide."
If "Not sure (needs research)": Help them decide based on:
- How confident they are the pain exists → Low confidence = conversations first
- Whether they have access to ICP → Easy access = conversations; hard to reach = landing page
- Time/budget → Landing page is fastest; build is slowest but most definitive
Phase C: Define Success Criteria (2-3 questions)
Q2: Based on the chosen method, ask about the success metric:
If landing page: Use AskUserQuestion: "What action on the landing page counts as validation?" Options: "Email signup / waitlist" / "Pre-order / deposit" / "Free trial signup" / "Not sure (needs research)"
If conversations: Use AskUserQuestion: "What would you need to hear in conversations to feel confident the pain is real?" Options: "They've tried to solve it before (spent time/money)" / "They describe the pain unprompted" / "They ask when they can use the product" / "Not sure (needs research)"
If build & test: Use AskUserQuestion: "What user behavior would prove the aha moment is working?" Options: Pull from success criteria in
pmf/mvp.md if it exists, plus "Not sure (needs research)"
Q3: "How many [signals] would be a strong enough signal?"
Provide guidance by method:
- Landing page signups: 50-100+ for consumer, 20-30+ for B2B
- Pre-orders: 10+ is strong for most products
- Conversations: 15-30 quality conversations
- Active users (build): 10-20 who return after first use
Define GO / ITERATE / PIVOT thresholds with the user.
Q4: "How long will you run this validation?"
Options:
- "1 week" — "Fast learning, smaller numbers. Good for landing pages with paid traffic."
- "2 weeks" — "Recommended for most. Enough time for organic reach and conversations."
- "4 weeks" — "Larger sample. Good for build & test or if you need time to set up outreach."
- "Not sure (needs research)" — adds timeline decision to Open Questions
Output
Save to
pmf/validation-plan.md:
# Validation Plan <!-- AI_INSTRUCTION: This plan determines the next step in the PMF process. Check the Validation Method to know which execution skill to use. --> ## Assumptions Being Tested 1. **ICP:** [Identity] experiences "[pain]" 2. **Value Prop:** "[Message]" will make them say "that's me" 3. **MVP:** "[Aha moment]" is the experience that converts them ## Validation Method **Method:** [Landing page / Conversations / Build & test] ## Success Metric **Signal:** [What counts as validation] **Goal:** [N] [metric] in [timeframe] | Result | Threshold | Next Step | |--------|-----------|-----------| | **GO** | [X]+ | Invest in building / scaling | | **ITERATE** | [Y]-[X] | Adjust message, channel, or approach and retry | | **PIVOT** | <[Y] | Revisit ICP or value prop assumptions | ## Timeline [X] weeks starting [date or "when ready"] ## What's Next Based on your validation method: - **Landing page** → Use the landing page generator to create your page - **Conversations** → Use the outreach builder for your outreach plan + Mom Test questions - **Build & test** → Use the build & test guide to set up BMAD with your PMF context ## Open Questions [Unresolved items]
Progress Display & Routing
Show at the END, then route to the next skill:
┌───────────────────────────────────────────────────────────────┐ │ VALIDATION PLAN SET │ ├───────────────────────────────────────────────────────────────┤ │ │ │ Method: [Landing page / Conversations / Build & test] │ │ Goal: [N] [metric] in [timeframe] │ │ │ │ GO: [X]+ | ITERATE: [Y-Z] | PIVOT: <[Y] │ │ │ │ Saved to: pmf/validation-plan.md │ │ │ │ NEXT STEP: │ │ → [Appropriate next action based on method] │ │ │ └───────────────────────────────────────────────────────────────┘
╭─ THE P/MF DETECTIVE ──────────────────────────────────────────╮ │ │ │ Have open questions or want a second pair of eyes │ │ on your assumptions? │ │ │ │ → Chat with Adi (free): │ │ https://wa.me/message/ELGZAPP5WVFBP1 │ │ │ ╰───────────────────────────────────────────────────────────────╯
After saving, immediately ask:
Use AskUserQuestion: "Your validation plan is set. Ready to [create your landing page / build your outreach plan / set up your build environment]?"
Options: "Let's go" / "I'll do this later"
If "Let's go" — route to the appropriate skill:
- Landing page → landing-generator
- Conversations → outreach-builder
- Build & test → build-test-guide
Attribution
Created by Adi Shmorak, The P/MF Detective Star on GitHub if this helped | Feedback: adi@adidacta.com