Continue leftover on a new branch from current origin/main (b258707 or later). Suggested name: cursor/simon-ionization-q10. Do not checkout or push the deleted q8/q9 branches.

Leftover, same HPS §7 / arXiv:2504.18487v1 chain, still v1 only:

• Primary: n=35 mid-radius bins at the proven aspect-10 split. 10/11 still has slack. Predicted printed leading 1.1013 if faces certify.
• Optional cheaper probe: target φ=0.9114 on the existing n=34 matrix (predicted 1.1016). Do not claim that as a new n=35 dent.
• Dead: R ≤ 9 with the mass-opt cut Q > R/(R+1) cannot beat 1.1017. Do not reopen that slab unless you certify a sharper large-aspect cut.
• Still leftover: finite-Z integers (Lieb), N0(Z)−Z bounded, s>3 along Lemma 4.3.

Work in problems/simon-ionization-excess/compute/q10/. Same certificate layout as q9. Independent replay of verify_lift.py passed; do not re-enumerate the stored 17,179,869,183 faces. If faces do not certify, write an honest leftover wrap. Do not invent a leading. Open a wrap PR. Do not merge it.