If the diode "burns" (melts) easily the goggles (as black plastic of the same kind), then they absorb well it's wavelength.
I have $3 red goggles which i tested on a 405nm Sony SLD3235FV diode sold on ebay.
I have accidentally pointed the diode into my eyes through the goggles. I saw it like emitting tiny red spot despite it's an UV diode. When I look at reflected from non-fluorescent material beam, I can not see it.
My eyes did not get hurt, but I wasn't running the diode at it's full power. Actually it was 5 promiles PWM at 90mA, because I needed to look at the spot to focus it and it shoud not burn the matte plastic I used to focus at.
5 promiles = 0.005% ( Fpwm=2kHz, Tpwm=2000, Ton = 10 ).
I think in general it could be said that with goggles that block out a large part of the spectrum (say pass red, block green and red) the pigment is throughout the lens and a scratch will not reduce the optical blocking much.
In case it blocks a narrow wavelength range (such as blocks 532 but passes 500 and 600 nm) there could be a dielectric coating, and a scratch could remove that entirely. This would probably be visible though.
I've never seen goggles of the latter type, but i believe they do exist, probably for use in specialist applications where good color vision must be maintained (military etc).
I agree with the above (there's a typo... they block green and blue]). Moreover I've heard that there exist prototypes of goggles with several selective coatings - for the most common wavelengths of LASERs to protect the military pilots. But I have not heard about their prices.
My $3 glasses block the green and blue light and pass only the red. But this greatly reduces the color vision.
There's another problem: to focus the UV beam, I need to see the spot. So the red ones are inappropriate for this task. I do this at low power with a pair of "retired" yellowish sunglasses which employ polarization filter also. They do not block 100% of the 405nm (neither do my Polaroid sunglasses) so I can work.