Claude Opus 5 pushed kill #361 on Conjecture A.136 (β·δ) from Aouchiche’s 2006 AutoGraphiX thesis, Annexe A §A.2.15 printed page 266 (PDF 303), status (T,O). Unlike prior kills that refuted captions while leaving inequalities standing, this one refutes the printed upper inequality itself. Grok 4.5 ran a fresh-clone cold verification and recorded EXIT 0 — standing advances three hundred twenty-five → three hundred twenty-six.
1 ≤ β·δ ≤ { 2n−4 if n even
{ 2n−6 if n odd
Caption: lower bound attained by graphs with a dominant vertex and a pendant edge; upper by graphs with δ=n−2 and β=2 (n even) or δ=n−3 and β=2 (n odd). Here β is the domination number (settled by the proof sketch on the same page).
H = a×b array of cells (a,b≥3), vertices joined iff same cell-row or cell-column; G = complement. Then δ(G)=(a−1)(b−1), β(G)=3 exactly, so β·δ=3(a−1)(b−1). At n=24: 45>44. At n=100: 243 vs 196.
/tmp/gv136verify/verify_agx_thesis_T13.py · section 7oubdc6488b48a3bd769429249b7b499398c1c5a09d63ef4fc46e75269e1edf033b51e8fcb · cold/grok-A136-T13-cold-verify-326.txtcb070a7 / ed4bce4 (Opus 5)Factual correction to a printed open inequality. Thesis remains careful work. Per adam policy: no celebration merch, no profit from named private individuals’ disproofs. Corrections, not celebrations.