Well, first of all, my apologizes for picking up old thread, but I’m right on subject and feel kind of guilty for not posting a bug report.
However, there is a good workaround for those color artifacts that I did not find elsewhere.
In the latest Enfuse 4.2 manual I have found extra parameters for color model:
(sorry preformatted text did not work here, pasting as image instead)
My parameters of choice were --blend-colorspace=cielab. No frenzy colors were found after using this parameter. Moreover, I extensively use Lab within Darktable for panorama postprocessing so this kinda aligns it. I also added the same parameter to Hugin engine and Holger’s HDR script into darktable. Feel free to ask for instructions where to paste them.
Combined with other parameters set properly it gives very good results 
Happy fusing everyone!
I guess Enfuse fails to identify colorspace properly and enforces RGB, which is not a good choice for luminosity blending. Please consider my last comment, what do you think about it?
Thinking back, @patdavid recommended using l-star (=cielab) in one of his tutorials. However, as @Morgan_Hardwood noted, that would not negate the alleged problem with ciecam.
Yep! Correct me if I’m wrong, EnfuseGUI is an old GUI that does not work on contemporary distros.
Today I tested the options “–blend-colorspace=cielab” and “–blend-colorspace=identity”. I was not able to reproduce the error with “–blend-colorspace=identity” from my post above, so both options seem to work for me so far.
at some point ‘identity’ stopped working for me and the artifacts started coming appearing again, so I opted for ‘cielab’ which was more stable. I guess you are right as I too read somewhere that enfuse forces RGB model