kelly-behavior-predictlisted

# Predictive Recommendation Analytics Desk ## Overview Use this skill as a Busabase-backed analytics dashboard for a **generic, brand-free** consumer booking/e-commerce product. It aggregates a fully deterministic mock user-behavior sample — 100 sessions across 5 session archetypes ("segments") — into a funnel drop-off view, a rule-based "predicted next action" per segment, and a prediction-accuracy backtest. Default behavior is AirApp-first. Unless the user explicitly asks only for explanation, ensure the mock sample exists (run the seed script below if the `sessions` Base is empty) and give the user the clickable AirApp URL (or the local preview URL when local preview is explicitly requested). Use chat-only mode only when the user says "chat only", "no UI", "纯聊天", "不要打开 UI", or similar. This is primarily a **dashboard** app type (read-mostly, no approval lifecycle). It carries exactly one narrow human-review surface: marking a segment's prediction rule "trusted" or "needs recalibration" with a note, written directly onto that segment's own Busabase record. That review never edits the rule, the dataset, or any live system — it is a review record only. ## App UI Screenshots <table> <tr> <td width="50%"><img src="assets/screenshots/overview.webp" alt="Analytics Desk overview"></td> <td width="50%"><img src="assets/screenshots/funnel.webp" alt="Analytics Desk segments"></td> </tr> <tr> <td><strong>Overview</strong><br>Overall funnel drop-off (browse → se