Fully agree. I guess there is still some honor and the other person would invite the original author as an author of the fixed paper, but it looks very snipable.
This is exactly what happened with Perelman's proof of the Poincare conjecture: https://www.newyorker.com/magazine/2006/08/28/manifold-desti...
But yes, if the proof is only missing easy plug-able gaps then likely a mathematician would reach out to the author so they could co-author the paper. Anything else would be rude.
I was more meaning in the case where the proof has some insight but not apply it in a way that actually disproves the Riemann Hypothesis.