Three different threes — the ℤ₃'s in the Standard Model are distinct

Recovered from the frozen log, where it was an in-run audit finding never filed (docs/CHARTER.md, the compression-loss finding).

Several "threes" show up in the Standard Model and its algebraic reconstructions, and the finding is that they are distinct — not one three wearing three hats. Conflating them is the error the audit caught:

• the gauge-centre ℤ₃ of SU(3) colour — isogeny / global-form data ([[CASE-03]]);
• baryon triality — a discrete ℤ₃ whose per-generation anomaly forces the generation count to be a multiple of three (Ibáñez–Ross type): a statement about the number of families ([[CASE-02]]);
• the S₃ of octonionic triality — the outer symmetry permuting the three 8-dimensional representations of Spin(8): the candidate kinematic source of "three" ([[CASE-06]]).

A centre (a subgroup), a discrete global symmetry with an anomaly, and an outer automorphism. They coincide numerically at three and are routinely merged, but they are different objects, and the "three families" question ([[CASE-02]]) turns on keeping them apart — which three, forced by which mechanism. A fourth "three" added to make the list feel complete is exactly the merge the audit warns against.