From build-options
Decide WHAT to build after a market is validated: generate divergent build options, score them with an independent judge panel into a weighted decision matrix, adversarially stress-test the top, recommend one (with kill criteria), render a Tufte decision matrix, and hand the winner to the `prd` skill. Use when the user asks "what should I build", "what are my build options", "which option should we build", "decide what to build", or after running `market-validation`.
How this skill is triggered — by the user, by Claude, or both
Slash command
/build-options:build-optionsThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Turns a validated market into a *defensible choice* among concrete build options. Runs the judge-panel
Turns a validated market into a defensible choice among concrete build options. Runs the judge-panel
method (generate → score → stress-test → choose). Pairs with market-validation (before) and prd (after).
Read references/decision-method.md and references/chaining.md first.
market-validation run,
but confirm the framing in Phase 0 before launching.Build a DecisionContext (references/chaining.md): chain from a prior market-validation run (read its
deck-data.json + brief) if available, else collect a short standalone brief. Then confirm with the user via
AskUserQuestion: budget/time horizon, risk appetite, and must-reuse assets. Present the **default criteria
references/decision-method.md) and let the user adjust (weights must sum to 1.0). Optionally
adjust the strategic lenses.You supply the context, criteria, and lenses; the script executes them (it does not invent them):
Workflow({ scriptPath: "<skill-dir>/assets/build-options-workflow.js",
args: { context: DecisionContext, criteria: [...], lenses: [...] } })
It generates one option per lens, has 3 independent judges score every option on every criterion, aggregates
to per-criterion {mean, spread, notes}, stress-tests the top ~3, and returns
{ options (with scores + adversarial), criteria, judgeCount, stressTested }.
decision-data.jsonWrite ONE canonical decision-data.json (mirror references/example-shiftmate/decision-data.json):
meta, context, criteria (with weights), options (the workflow's output verbatim — raw scores means;
do NOT hand-compute weighted totals, build_matrix.py does that), recommendation (your call: winner +
rationale + grafts + confidence), and killCriteria. If the workflow's top option was killed in the
adversarial pass, do not recommend it without re-scoring.
python3 <skill-dir>/assets/build_matrix.py decision-data.json --out <out-dir> --pdf
Computes weighted totals + tie analysis via weighting.py and renders the Tufte matrix (<slug>.html;
+ .pptx if python-pptx; + .pdf with --pdf + chromium). If tie.isTie, lead with the deciding factor.
prdInvoke the prd skill with a feature description = the winning option's thesis + mvpScope +
businessModel + key constraints (see references/chaining.md). It writes /tasks/prd-*.md; the tasks
skill then turns that into prd.json. Surface all artifacts to the user (SendUserFile).
decision-data.json, <slug>.html (+ .pdf/.pptx if produced), and the PRD for the chosen option.
build-options-workflow.js) is syntax-checked only — its first real invocation is its
proving run. Confirm args (context/criteria/lenses) flow into the agents and the returned
{options, criteria} shape matches what Phase 2 consumes before trusting the output.prd handoff is described, not yet exercised — confirm the PRD lands in /tasks/ the first time.python3 -m pytest <skill-dir>/tests -q — unit-tests the weighting math (aggregation, weighted sum with
reg_risk inversion, tie detection) and smoke-tests the matrix render (golden + minimal). Run after editing
weighting.py or build_matrix.py.
references/decision-method.md — rubric, math, tie/kill rules.references/chaining.md — read market-validation output; standalone brief; the prd handoff.references/example-shiftmate/decision-data.json — worked example (ShiftMate).Provides UI/UX resources: 50+ styles, color palettes, font pairings, guidelines, charts for web/mobile across React, Next.js, Vue, Svelte, Tailwind, React Native, Flutter. Aids planning, building, reviewing interfaces.
Fetches up-to-date documentation from Context7 for libraries and frameworks like React, Next.js, Prisma. Use for setup questions, API references, and code examples.
npx claudepluginhub craigm26/founder-skills --plugin build-options