Kinetic demo follow-up: the N-window mailbox beat + gesture-drive decision

Context

Operator narrowing decision (2026-07-21, in-session ratification of Euclid's PR #15644 review fork): #15631's AC3 was amended to descope the N windows (+mailbox) footage beat here. The mailbox pop-out path itself is already verified (AC1 receipts on #15631: reveal → Pin → scripted tear-out; vessel born mid-gesture, tearOutPanes.operator committed, the live pane adopted inside the vessel). What remains undelivered is the mailbox film beat and its InteractionService-driven gesture witness — gated on the operator's #15536 polish freeze and the #15635 vessel-death gap.

Live latest-open sweep: checked the live tracker (created-desc, 12 latest) at 2026-07-21T18:50Z; no equivalent. A2A in-flight sweep: none (the recapture is frozen team-wide per Emmy's capture-hold). KB sweep: #15252 (the workstation five-beat capstone) is the sibling shape on a different surface — not a duplicate.

The Problem

The kinetic QT beat currently proves the detail's click-vessel loop (pop-out → glide → re-integration). The ticket's original ambition included the operator-mailbox as the second simultaneous popup (N windows) plus the mid-gesture live-pane transfer driven through InteractionService — the strongest form of the object-permanence proof. Today the mailbox's only window path is the scripted tear-out gesture (no cockpit-side semantic executor exists for it), its vessel-death reintegration strands the pane (#15635), and the app surface is under the operator's polish freeze (#15536 / Phoebe's #15649).

The Architectural Reality

• The mailbox pop-out path works: reveal → Pin → bundled simulate_event tear-out (receipts on #15631). The drive needs no new machinery for one-off captures.
• #15635 (vessel-death stranding) sits in the #15239 G4 leaf's scope — the reintegration-on-close must land before a reattach-home mailbox beat can be filmed honestly (reload-restore is the current workaround).
• A cockpit-side scripted gesture executor (the DemoBCrossWindowDragNL.executeCrossWindowStep pattern, cockpit flavor) is the open design question: the #15631 ticket forbade new pop-out machinery; a drive executor is choreography, but the call belongs to the dock-machinery owners (Clio's #15243 matrix lane defines the gesture contract this would ride).
• The #15536 / #15649 polish arc is the operator's gate for any app-surface re-shoot.

The Fix

After the polish gate and #15635 land: one clean capture adding the mailbox beat to the delivered kinetic choreography — reveal → Pin → scripted tear-out as the second popup (mid-gesture birth), hold N windows live (detail vessel + mailbox vessel + main), then re-integration of both (detail via reattachAgentDetail; mailbox via its G4-fixed reintegration path or reload-restore documented). Evaluate then whether the scripted drive stays raw simulate_event bundles or earns a cockpit-side executor (decision recorded with Clio's matrix lane).

Acceptance Criteria

• Footage shows 3 simultaneous live windows (main + detail vessel + mailbox vessel), all updating.
• The mailbox vessel is born mid-gesture from the scripted tear-out (the signature boundary crossing).
• Both panes re-integrate same-instance (mailbox path per #15635's fix or the documented reload-restore).
• First/mid/final decoded frames verified before handoff (the #15631 pass6 lesson).
• The executor-vs-raw-drive decision is recorded with the matrix lane.

Out of Scope

• The detail click-vessel loop (delivered in #15631 / #15644).
• New pop-out machinery (the original #15631 constraint stands).
• #15635's fix itself (G4's scope).

Related

#15631 (parent narrowing) · #15644 (the regeneration substrate) · #15635 (reintegration gate) · #15243 (gesture contract) · #15252 (workstation five-beat sibling) · #15536/#15649 (polish freeze)

Origin Session ID: 557fd3c7-7307-499b-9c35-d07fe1c2efcd

Retrieval Hint: mailbox film beat N windows tear-out second popup gesture executor FleetCockpit reintegration #15631 descope