cleanup before we spcify and implement more clearly

agent-tasks/prepare-beta-plan.md

@@ -0,0 +1,35 @@
+# Agent Task: Prepare Beta Plan
+
+## Goal
+Draft a small, controlled beta plan for a named prototype.
+
+## Context
+`whynot` uses closed betas to learn quickly without implying production maturity. Beta rules are defined in `BETA_MODEL.md`.
+
+## Inputs
+- A prototype card from `prototypes/`
+- `BETA_MODEL.md`
+- `templates/beta-plan.md`
+
+## Allowed Changes
+- Create or refine a beta plan in `betas/`.
+- Clarify the learning question, target audience, entry criteria, and feedback method.
+- Recommend the most appropriate beta type.
+
+## Forbidden Changes
+- Do not announce a public beta.
+- Do not promise production readiness.
+- Do not add payment or investment mechanics unless explicitly approved.
+
+## Expected Output
+- One beta plan with a clear learning question, beta type, entry criteria, feedback method, and end condition.
+
+## Acceptance Criteria
+- The plan names the prototype and audience.
+- The plan explains how feedback will be captured.
+- The beta is small and reversible.
+
+## Human Approval Required For
+- Public launch language.
+- Paid beta mechanics.
+- Cross-repo implementation commitments.