I guess that adding filters and a stepper makes the idea too complex.
I didn´t know the leds drifted significantly with juncture temperature. Sounds very reasonable of course. I wonder if during a normal exposure time this would be a problem. In any case maybe a "preheat" phase (somehow shutting-off the enlarger out - with a stepper motor and some blades hehe) might mitigate the problem.
For what I read so far, it seems that the main problem is the green contents of blue leds. This is indeed a problem for contrast but for color, rather I was thinking: unless the printing needs no green exposure at all, it could be compensated with less exposure from the green led. This could be easily be solved with software once the system is calibrated... of course calibration would need some extravagant instrument like an spectrometer.
I also read (I think PE was saying somewhere) that there is sensitivity in the paper outside the leds range, and thus the paper will not work as expected.
Then I wonder: which kind of leds do ra-4 digital printer use??