The c.p. needs to be significantly behind the c.g. in order to be stable either in boost or in glide.
The center of lift (c.l. for lack of better shorthand, but don't confuse it with the lift coefficient cl) needs to be at or very near the c.g. for glide.
In RG, the trick is to reconfigure something from the boost mode to get to a stable glider. You can rotate the wings, you can slide the wings, you can slide the expended motor casing, you can deploy an elevator tab on the horizontal tail . . . there are as many ways to do this as you can think up. It still has to be stable as a regular rocket during the ascent and then change to be stable as a glider for the descent.
Maybe you can find something to get started with on the NAR website?
https://nar.org/competition/plans/rocketglide.html