Classification: exact geometric identities and a polynomial countermodel to a symmetry/count inference. No numerical zeta experiment or RH result.
The linked Zhao handout was read in full, with its relevant diagrams checked. Its original bytes are not republished here; source URL/hash and precise scope are in sources.json. The earlier local Maxima copy did not contain the PDF. Exterior similarity is derived directly, not attributed to the handout.
See argument.tex, checks.mac (standalone ordinary Maxima, no Clifford
package), and ../../audits/tangent-circles-and-rh-2026-09-23-v1.html.
Run the worksheet with maxima --no-init --very-quiet --quit-on-error
--batch=evidence/ns94_circle_inversion/checks.mac; all 23 exact checks pass.
The figure is an illustration of exact rational-coordinate control points,
not a computation of zeta zeros. The source plot can be regenerated with
make_figure.py; floating plot coordinates do not enter the proof.
The formal note separates inversions in each circle, both orders of a successive pair, and inward/outward interpretations of the 0/1 variant. The positive power-product statistic is only a location detector; its vanishing at every height is an unproved RH-equivalent target. No general geometric route is excluded. Review is by the author only.