I tested a filter from a CNI laser, and it takes 6% off of my already filtered CNI...
But you probably don't have to worry about IR. There is very little left due to the way the laser is built, and the little that is left is highly divergent... It can be somewhere in the vicinity of a stronger IR LED, in power as well as spread..
You should be much more worried about the green beam hitting your eyes, or an unexpected reflection.
IR was much more of a problem, when the crystals had a much lower efficiency, and a more powerful pump had to be used. The IR vs. green ratio was much higher, and the IR was collimated, before it entered the crystals.
In these lasers, the IR enters the crystals uncollimated, while the green comes out very narrow.
This is then expanded with the expamnder lens, along with a little IR, while the nut blocks most of the IR.
The collimator nut again mostly collimates the green, a little IR, while the nut holding it blocks most of the remaining IR.
Then, if the laser has a tube extending further up from the collimator, ending in a narrow aperture, this again mostly lets the green pass while blocking most of the little IR that is left.
When i was taking green lasers appart and putting them back together, i observed the amount of IR decreasing at each of these steps.
At the expander there is a huge IR glow, at the collimator it is smaller, and at the apperture, it is MUCH smaller. If produces maybe a lens flare, but doesn't blind the camera..
When i measured my lasers with a known transmissiveness IR filter, and then added the percentage back to the green, that it takes away, i got the same result as if i measured directly.. The little IR left is hardly measurable.
Otherwise, yes it is safe to glue the IR filter into the cap from the inside. Just be careful, that you don't get glue on the middle. But most likely it is not necessary, and will decrease your green output. Not visually noticeable, but measurable.