Create a consumer-prefill umbrella handoff.

Accepts the fullest consumer umbrella application or estimate you can assemble, stores it at Coverage Cat's one-shot review checkpoint, and returns the handoff uid, a rotating intake_access_token for direct API follow-up, a short-lived resume_url browser fallback, and a coarse public polling_token. On a cold start, call this before you question the shopper. On the initial consumer umbrella handoff, prefer sending only intake; add top-level field_estimates only when you are certain each field already matches a current umbrella review field path. Use canonical consumer umbrella keys such as address.street, address.city, address.state, address.zip, license_state, license_number, approximate_asset_value or net_worth_numeric, desired_coverage_limit, required_coverage_e_limit, and vehicle_count or umbrella_details.motorized_vehicle_count for umbrella counts. Avoid top-level street/city/state/zip, drivers_license_state, drivers_license_number, net_worth, or desired_umbrella_liability_limit. Do not send vehicles rows before submission. Use vehicle_count or umbrella_details.motorized_vehicle_count for the household vehicle count; actual vehicle details are collected later only if a selected offer requires them, such as a Markel bind follow-up. Optional consumer sandbox: set top-level sandbox: true only on this first prefill call when you want a fully synthetic shopper rehearsal. That same consumer sandbox uid can then keep using /api/intake/{uid}/issues, /api/intake/{uid}, /api/intake/{uid}/select, /api/intake/{uid}/bind, /api/intake/{uid}/attach, /api/intake/{uid}/payment, /api/intake/{uid}/policy, /api/intake/{uid}/service, and /api/consumer/status with synthetic quotes, uploads, payment credentials, policy outcomes, and service-request results only. Until a later follow-up response shows quote progress, the application is not submitted yet. Coverage Cat may answer with a staged review_summary.review plus a bundled review_summary.next_question so the assistant can collect the shopper-owned details it still needs, such as full name, email, phone, full address, birthday, marital status, driver's license state, driver's license number, net worth, and any corrections, in one turn instead of drifting into a field-by-field questionnaire. When that happens, tell the shopper explicitly that Coverage Cat has not submitted the application yet, ask for those remaining items in plain English, and render the gathered details plus remaining items as short labeled bullets or sections rather than a prose paragraph. If any shown value is estimated, mark each estimated bullet label or value with *, include the short note * = estimated once above and once below the list, and do not prefix every estimated line with [Estimated]. If the staged review shows a wrong estimated umbrella count from context, patch the canonical umbrella_details.* count field before you show the final review. Coverage Cat will not resume or expose a prior consumer umbrella draft based on email alone. Continue an existing consumer draft only through the issued uid plus the current hidden MCP session or latest intake_access_token handoff. If the shopper replies with a birthday in MM-DD-YYYY form, normalize it to YYYY-MM-DD before the follow-up patch. Successful direct follow-up responses rotate a fresh intake_access_token; always reuse the newest token you have. Credit consent is not required here — Coverage Cat collects the soft-credit pull authorization after the user chooses an offer.