playbooks/zero-input-defaults.md

# Zero-Input Defaults

Use this playbook only when a primary ForgeMind journey is invoked without a user-supplied outcome. State the derived goal in the first response and label it as a zero-input default. A supplied user outcome always overrides this playbook.

## Compass

Inspect the project, Git state, and prior ForgeMind records. Select the path with the highest provable leverage: Spark for new radical directions, Evolve for an existing application, Venture for commercial validation, Council for an unresolved decision, Ship for delivery, or Leap for autonomous end-to-end work. Do not ask routine questions; explain the selected path, derived goal, assumptions, and next visible result.

## Spark

Develop five radical AI product ideas for this project. Do not suggest small improvements or generic chat features. Identify the product purpose, user flow, needless steps, recurring manual work, and strongest differentiator. Every idea must eliminate a complete workflow, make AI central, state a 10x hypothesis, reduce or remove UI, provide USP and moat, metric, kill condition, and smallest buildable MVP. Rank impact, differentiation, feasibility, risk, and time to MVP; select the strongest reversible idea.

## Evolve

Analyze the existing application before proposing changes. Identify its product purpose, target user, decisive job, current flow, unnecessary steps, technical constraints, available data, and integrations. Create five AI-central transformations that replace a complete workflow. For each provide USP, 10x hypothesis, metric, kill condition, MVP, risk, and rollback. Select the strongest reversible transformation, then define a build-ready component structure, AI core, data flow, feature flag, tests, and success measurement. Repository findings are project evidence, never evidence of customer demand.

## Venture

Build a complete venture case for the strongest available opportunity: customer problem, segment, alternatives, differentiation, USP, willingness-to-pay hypothesis, pricing, route to market, and validation needs. Calculate market chance from pain, frequency, reach, differentiation, feasibility, and evidence. Produce conservative, base, and upside scenarios covering reachable accounts, price, revenue, margin, CAC, churn, build cost, run cost, break-even, twelve-month net effect, and sensitivity. Return Go, Validate, Park, or No-Go. Do not invent sources, market size, customer evidence, or forecasts.

## Council

Find the most consequential unresolved product, architecture, or delivery decision. Consider product value, customer demand, technical feasibility, risk and rollback, and a contrarian view that asks what should be deleted or not built. Record facts, assumptions, dissent, decision, owner, metric, kill condition, safety boundary, and next action. Prefer the smallest reversible validation when uncertainty remains.

## Ship

Select the latest ForgeMind bet, or otherwise the smallest measurable MVP with the highest user leverage. Create a delivery contract with goal, non-goals, acceptance criteria, minimal architecture, components, AI and data boundaries, feature flag, rollback plan, metric, kill condition, and risks. Implement the smallest vertical slice rather than only planning. Cover functional tests, critical user flow, loading, empty, error, success, keyboard, narrow viewport, and accessibility states. Inspect the final diff and return evidence-backed Go/No-Go. Continue without routine questions and stop only at a hard safety boundary.

## Leap

Start the complete autonomous product mode. Analyze the current application when one exists; otherwise create a disruptive opportunity from project context. Produce five radical AI-central options that eliminate or replace full workflows. Reject conservative feature additions and dashboard-first concepts. For every option provide 10x hypothesis, USP, moat, metric, kill condition, MVP, risk, and rollback. Assess market chance and conservative, base, and upside business-case scenarios. Choose the strongest reversible bet plus a credible contrarian alternative, then implement its scoped MVP with architecture, components, AI core, data flow, feature flag, functional and GUI tests, accessibility, risk checks, rollback plan, and release decision. Treat missing market, ROI, customer, and test evidence as assumptions or gaps; never as facts.