mvp-type-selectorlisted

# MVP Type Selector An MVP is not a small product. It is an experiment that answers one question at the lowest possible cost. This skill picks the MVP type that tests the riskiest assumption, names the pass/fail threshold before anything ships, and refuses to let "build" be the first step of the loop. ## The Rule The right MVP is the one that produces the required learning for the least build effort. If two types answer the same question, pick the one you can kill fastest. Speed of learning beats fidelity of product. ## Modes ### Select mode (default — user describes an idea or hypothesis) 1. Read `references/mvp-type-catalog.md` before recommending anything. 2. Extract the hypothesis. If the user has not stated one falsifiable hypothesis, ask for it once — an MVP without a hypothesis is just a rushed product. 3. Identify the riskiest assumption the MVP must test: demand ("will anyone care?"), value ("will they use it enough?"), or willingness to pay ("will they give up money/time/reputation?"). 4. Recommend one MVP type with: why this type, what it tests, what it deliberately does not test, estimated cost, days to first learning, and the kill condition. 5. Output the test plan from `templates/mvp-test-plan.md` with every bracket filled. ### Plan mode (user has already chosen a type) 1. Confirm the type exists in the catalog; if the user invented a hybrid, decompose it into the catalog types it combines and say so. 2. Fill the test plan. The pass/fail t