nx_seat_gate.nx

nx_seat_gate.nx -- makes nx_seat's OWN proof reachable by the ship loop. THE GAP THIS CLOSES, measured 2026-08-22: nx_seat IS gated -- `nx_seat selftest` runs 29 gv_check teeth and prints "NX-SEAT-GATE passed 29/29 verdict=GREEN". But that proof lives behind a VERB, and osl_gate_resolve resolves only gate BINARIES (<target>_gate by convention, or a declared row in knowledge/organ_gate.conf whose value is a gate NAME). So every ship of nx_seat reported `prove=NO-GATE-FOUND UNPROVEN` while a 29-tooth gate sat one argument away. *A PROOF THAT THE SHIP LOOP CANNOT REACH IS, TO THE SHIP LOOP, A PROOF THAT DOES NOT EXIST. WHY A THIN GATE AND NOT A RESOLVER CHANGE: organ_gate.conf already treats SPACE as a field separator, so encoding "<binary> <verb>" there needs a NEW SYNTAX DECISION, not a mechanical fix -- and the resolver is the path every organ in the estate ships through. Satisfying the existing convention costs one small binary and changes nothing anyone else depends on. *WHEN A CONVENTION AND A SUBJECT DISAGREE, THE CHEAPER AND SAFER MOVE IS USUALLY TO SATISFY THE CONVENTION, NOT TO REDESIGN THE CONTRACT EVERY OTHER CALLER SHARES. This is a PASS-THROUGH, deliberately: it does not re-implement or re-judge anything. It runs the real selftest and requires BOTH that the process exited 0 AND that its verdict line says GREEN -- because an exit code alone cannot distinguish "green" from "the binary was replaced by something that exits 0", and a string alone cannot see a crash after printing.