Dark Matter Scaffolding: Graph Regularization and the Prevention of Local Big Rips

Paper 97 does not claim to have detected, identified or excluded any dark matter particle, to have derived any real galaxy's rotation curve from baryonic physics, or to have shown that the universe allocates hidden nodes. It shows, in one reproducible experiment on a sampled graph, that (i) a rotating disc whose supplied mass is below the requirement M_req = v_obs² r/G tears from the rim inward, with a deterministic per-node detachment time 1/[κ(τ − 1)] that reproduces 3,128 measured detachment times across almost two decades and seven scale lengths from one parameter-free formula; (ii) the naked galaxy loses 78.2% of its mass by t = 20 against a predicted 78.7%, and its remnant rotates at 0.299 of the observed speed; (iii) a hidden allocation of f_c = 1 − M_vis/M_req = 0.590 keeps 100% of the disc, and f = 0.62 loses not one node; (iv) thirteen runs across the allocation reproduce the parameter-free prediction to within 2.4 points in every cell, with a sharp corner at f_c and a flat plateau above it; and (v) the deficit allocation M_dark = M_req − M_vis has dM_dark/dr → v_flat²/G exactly, giving 0.4761 against 0.5000 at the truncation edge and the projected density slope −0.78 trending to −1 — an isothermal halo produced by a rule rather than fitted by a parameter. The profile is not NFW: the local-deficit rule cannot produce a cusp, because a cusp would require oversupplying the centre relative to the local demand. What the paper claims philosophically is only this: if the profile of the missing mass is what a maintenance budget looks like, then the question "what is dark matter made of" may have the same status as "what is a regularisation term made of".