Data and code for: Historiogenic universes: an operational census of the SO(10) flavor lattice

A companion paper finds a quarter in the benchmark fits of minimal SO(10) and traces it to a democratic-clockwork chain whose subleading eigenvalue $1/m$ quantizes the flavor expansion parameter. The chain has two integer dials, the family number $N_f$ and the register dimension $m$, and turning them defines a lattice of neighbor universes inside one fixed theory, ours at $(3,4)$. This paper walks that lattice. Its seven computed cells share one symmetry-breaking pattern and differ in how the six neighbors die. The chain fixes the channel structure, the family-depth spectrum requires minimality, and on the equal-weight diagonal a no-wrap charge window selects the same charge spectrum. The result is spectrum-level: site occupancy and chain length need a separate premise that the certified register observables cannot supply. The later of two epoch classes arrives at weak freeze-out. In the exact cells $(4,5)$ and $(5,6)$ the charged nucleon ends heavier than its neutral partner in almost every conditioned draw on the primary convention, and never far below nineteen in twenty across the tested tolerance-3 conventions. One branch escapes. In roughly two in five to one in two of the calibrated draws the ground state carries charge $-1$, and its charge-conjugated hydrogen passes every computed gate at a probability the same criteria bound near one in seven. No published two-baryon calculation covers its split pion multiplet. Exact spin-flavor algebra reduces the total exchange coupling to about a third of the ordinary deuteron channel, and a calibrated leading-order model leaves the residual binding contact-dominated, so a lattice calculation with non-degenerate light quarks would decide the branch. Two cards live long enough to simulate. The exotic $(5,6)$ cell forms a dark web but no stars in one adiabatic run. Across eight mirror $(4,5)$ pairs that share initial conditions, the mirror holds about four percent more halos above the analysis floor and about nine percent less supercluster-scale tracer volume, the sign holding in every pair. Both exhibits are descriptive, dark-matter-only, on encoded cards. The $N_f = 0$ row is a pure-gauge archive: dynamical states, no demonstrated write channel. Matter, cosmology, and the register criterion leave $(3,4)$ as the only historiogenic cell on the computed table, holding stable matter, a nonzero flavor-sector T-asymmetry marker, and persistent register deposits at once. That is a convention-conditional algebraic characterization, not a derivation of a thermodynamic arrow, a dynamical selection, or an anthropic claim.