cleanup before we spcify and implement more clearly

BETA_SIGNAL_INTAKE.md

@@ -0,0 +1,59 @@
+# Beta Signal Intake
+
+## Purpose
+
+This playbook defines how a `whynot` idea moves from prototype intake through beta, signal review, and promotion decision.
+
+## Flow
+
+### 1. Intake and Triage
+
+- Capture loose ideas in `inbox/`.
+- Use `PROTOTYPE_PIPELINE.md` Stage 0 and Stage 1 to decide whether to park, reject, merge, or create a prototype card.
+- If the idea deserves structured work, create a prototype card from `templates/prototype-card.md` in `prototypes/`.
+
+### 2. Prototype Readiness
+
+A prototype is ready for beta planning when it has:
+
+- a named audience;
+- a learning question;
+- a smallest useful test;
+- an expected signal;
+- a plausible promotion target if the signal is positive.
+
+### 3. Beta Planning
+
+- Use `BETA_MODEL.md` to choose the beta type and confirm entry criteria.
+- Create a beta plan from `templates/beta-plan.md` in `betas/`.
+- Keep the beta small, explicit, and reversible.
+
+### 4. Signal Capture
+
+- Record feedback and market evidence in `signals/`.
+- Use `MARKET_SIGNAL.md` and `templates/signal-record.md`.
+- Preserve null, weak, and contradictory signals rather than rewriting history.
+
+### 5. Promotion Decision
+
+- Review prototype state, beta notes, and signal strength together.
+- Draft a decision from `templates/promotion-decision.md`.
+- Record the final promotion or non-promotion outcome in `DECISIONS.md`.
+
+## Required Artifacts
+
+| Stage | Required artifact | Location |
+|---|---|---|
+| Intake | raw note or captured idea | `inbox/` |
+| Prototype | prototype card | `prototypes/` |
+| Beta | beta plan | `betas/` |
+| Signal | signal record | `signals/` |
+| Decision | promotion decision record | `DECISIONS.md` and optionally a filled copy of `templates/promotion-decision.md` |
+
+## Seeded Agent Tasks
+
+- `agent-tasks/triage-prototype-intake.md`
+- `agent-tasks/prepare-beta-plan.md`
+- `agent-tasks/review-signals-and-draft-decision.md`
+
+Use these when the repo needs a repeatable agent prompt instead of ad hoc instructions.