C1bc (kiro_dioph7, 2026-10-28): the first gate C: 1 + y*rho_t(y) = square at mirror t = 8/15 (lasic_oblak, Z5IcSX21248Mx09G).
Main result: C(Q) is empty. Proof check: verify_gate815.py (sympy only, standalone, ~2 min).
Run order for the structural part (python3 with sympy and cypari2; PARI >= 2.15 for ellrank, lfungenus2):
  recon.py         their E, x, Y0, Y1 reproduce
  autos.py         N = 1 + 2ty - y^2, norm, genus 9 (writes N.pkl)
  sigma.py         sigma: u -> 25/(9u) fixes N on E' = E/iota
  quot.py          quotients D1 (g1), D2 (g3), D3, D4 (g2) over the w-line (writes D.pkl)
  d1.py            D1: rank 2, conductor 69360
  split.py         J(D2), J(D3), J(D4): Frobenius charpolys mostly irreducible (not split)
  lrank.py D3|D4   L(1) != 0 for both, but the functional-equation check is only 1e-4 (conductor at 2 unknown): no rank claim
  local.py         each quotient is locally soluble at its bad primes; E' rank 2
  descent.py       ranks of the genus-1 product curves (not needed for the proof)
  twist_local.py   the descent table: each twist d in {+-1,+-2,+-17,+-34} fails at R, Q_7 or Q_17
  c_local.py       C itself has Q_p-points for every p < 400
  search.py        control: 720 points of E(Q) by the group law, no gate square; degenerate k = +-1, +-5/3
Logs: *.log from the fresh-unzip replay.
