I don't have the specs for your angle blocks handy, nor do I specialize in mechanical calibrations. But I do have a couple of thoughts in general metrological principles.
1. Unless there is a compelling reason, I recommend always certifying your standards to manufacturer's specs. Sounds like you know what they are anyway.
2. Sounds like your standard has "apparently" appropriate specs for what you want to calibrate with it (with the qualifier that this is not my expert area).
3. Something sounds questionable when an outside lab is trying to "sell" you on certifying your standard at less than it's designed specs.
If I were speculating, I would guess it possible a vendor may not have adequate uncertainty to correctly calibrate to its full specs. So in the interest of still keeping the business, suggests calibrating to lower specs.
If the unit cal'd by the standard is +/-0.003, it is not desirable to have the standard certified only to +/-0.0025. You are barely better than 1:1 uncertainty ratio.
My vote is to find and use an adequate vendor that has correct capability to certify your standard to it's full specs (especially since it appears you need full specs to adequately calibrate your units).
Do what's needed to assure risk to products at your company is minimized.
That's my two cents. Hope that is of some help.