Anyway, any feedback appreciated, especially if all thing is absolutely pointless.
An interesting fact (thanks, Dave!) was brought to my attention - lack of initrd "support"... I personally try not to use initrd at all, so I totally forgot about it. It will be just additional flash image and extra ATAG generated by the script. But it looks like the script need some proper option parsing after all...
There is also a question about the FDT. I know little about the boot contract in this case, but I believe there won't be much problem with my "method". But this is something that the Boot Monitor would have to do as well, if it is supposed to boot the kernel - valuable feedback :-)
Regarding initrd, I'll upgrade the script, do some testing and update the description next week.
Cheers!
Paweł