Interesting! An excellent demonstration of the situation, Guillermo. Thanks. It makes sense that the red channel would be the last to be clipped in a blue sky, unless it were a sunrise or sunset.
I guess someone at Adobe has decided that a reconstruction of a blown sky that leans towards cyan is more acceptable than one which leans towards magenta.
The Adobe software may not be as intelligent as you think. Rather than identifying a blown sky, ACR may have merely identified a clipped area. When reconstructing blown areas, ACR pushes blown areas to neutral. The red channel is the only one intact in the photo, so ACR may have added blue and green, resulting in a cyan cast.
DCRaw has various highlight recovery options. See Guillermo's post
on highlight recovery and look at the reconstructions of blown facial highlights in his example. ACR pushes them towards white, whereas DCRaw looks at the adjacent colors in determining the reconstruction.
ETTR is a good thing, but when one pushes it to extreme, clipping and data loss will occur. The prudent photographer avoids data loss rather than depending on implementation of highlight recovery in software. With today's high performance sensors, a bit of underexposure can be tolerated if bracketing is not feasible. Since blown highlights are often relatively neutral in color cast, highlight recovery can often rescue inadvertently overexposed images. However, I think that successful reconstruction has more to do with the SNR than the abundance of levels in the highlights as claimed by some experts. Shot noise and rendering of the raw file will usually be sufficient to dither the image and prevent banding.