Assumption Testing: AI Skill for Product Management
Assumption Testing is an AI skill that tests the beliefs behind an idea, not the idea itself. It finds the riskiest assumption, phrases it as a falsifiable hypothesis, and picks the cheapest method that answers it.
npx skills add Uxcel-Lab/product-skills --skill pm-assumption-testing
What this skill does
Helps an AI assistant test the assumptions an idea depends on, instead of asking users if they like it.
- Decomposes an idea into the beliefs that must be true for it to succeed.
- Covers five assumption types: desirability, viability, feasibility, usability, and ethics.
- Ranks assumptions by importance and certainty, then tests the riskiest first.
- Phrases each assumption as a falsifiable hypothesis with a metric and a threshold.
- Sets pass or fail criteria before the test runs, so results can't be reinterpreted.
- Picks the cheapest method that answers the question, from interviews to painted-door tests.
- Asks about past behavior, not hypothetical intent.
- Documents each result and the next step: confirm, adjust, or change direction.
When to use it
Use it when you need an AI assistant to:
- validate a product idea before building;
- de-risk a feature or a bet;
- plan discovery experiments;
- turn assumptions into testable hypotheses;
- design an MVP or painted-door test;
- decide how much evidence a decision needs;
- critique an existing validation plan.
Decisions this skill helps you make
| Decision | Options |
|---|---|
| Low-fidelity tests | Surveys, interviews, and paper prototypes for problem and desirability checks |
| Demand tests | Painted-door, landing-page, and smoke tests that measure real behavior |
| Medium-fidelity tests | Clickable mockups, concierge MVPs, and Wizard of Oz before writing code |
| High-fidelity tests | A/B tests, feature flags, and live betas for already-validated ideas |
| Qualitative vs. quantitative | Qualitative for why, quantitative for what, cross-checked against each other |
| Sample size and duration | About five users for usability bugs, larger samples for conversion claims |
How the skill works
1
Establish the stakes
The skill identifies the decision, its reversibility, and the leap-of-faith belief behind the idea.
2
Build the hypotheses
It decomposes the idea into assumptions, ranks the riskiest, and phrases each as a falsifiable hypothesis.
3
Match the method
It picks the cheapest test that answers the question, with pass or fail criteria set up front.
Try it with
Related skills
1.0.0
MIT
Product Management
Processes
