Curing cyanoacrylate emits fumes which can deposit a white substance on nearby objects, especially if they have any oil on them. It can and has been used by law enforcement to reveal fingerprints. But it's not corrosive.
Most hardware-store silicone produces acetic acid fumes as it cures. Which can be corrosive, like any acid; especially if trapped in an electronic enclosure. I know there's specialized silicones that don't release acetic acid, which can even be used in direct contact with electronics to totally encapsulate them; but I haven't used them and can't make a recommendation.
However, since you have an outer housing which screws on before any of these substances are applied, I'd think you can use anything that provides an adequate seal.
And as for the sensor cap, I think it's just a dust cap. I've never heard mention of any need to keep a photodiode in the dark most of the time.
Oh, and Mistergreen came up with an idea in the DIY PAR meter thread that adding a purple filter gel to reject some of the green, while passing a higher proportion of photosynthetic red/blue, might give results closer to that of a real PAR meter. The filter I see on your LUX meter appears green, doing the exact opposite of what you want. I doubt it's a UV filter, as photodiodes have poor response to that. They are however very sensitive to IR, and if it does remove any of the non-visible spectrum; that would be it.