Color blindness is a design constraint, not an edge case
Around one in twelve men cannot distinguish some of the color pairs in your interface. That is not an accessibility footnote, it is a larger group than Safari users.
Tag
1 article on color blindness.
Roughly 1 in 12 men and 1 in 200 women have some form of color vision deficiency. The common forms are not "seeing in grayscale" — they are a shifted or missing response in one of the three cone types, which collapses part of the color space onto itself. Deuteranomaly and protanomaly, both red-green, account for the large majority.
The practical consequence is narrow and specific: hue stops being reliable as the only carrier of meaning. A red "error" and a green "success" that differ only in hue can land within a few ΔE of each other for a deuteranope. The fix is almost never to avoid red and green — it is to make sure hue is never doing the work alone. Add an icon, a label, a difference in lightness. Lightness survives every form of CVD intact.
Simulation is the cheapest possible check. The transforms here operate on LMS cone responses rather than crudely rotating hue, so what you see is a reasonable approximation of the actual perceptual collapse rather than a filter that looks vaguely wrong.