Approach for pathfinding along cover?

Curious if anyone has suggestions for how I could approach making an AI pathfind along cover. I also still need to be able to make the AI pathfind the shortest path (not using cover) in certain cases. When using cover it would be the “safest” path, basically pathfinding through points that are marked as being out of sight resulting in them having low cost. I’m using a NavMesh that defines the walkable space, and could see myself adding a PointGraph with points evenly spacaed along the edges of NavMesh to define paths only where cover is placed.
Will two graph types play nicely together?
Instead of using a pointGraph, could I place my own type of AI Nodes which the pathfinder prioritizes and builds a path along using the NavMesh graph?

Basically what you want is penalties on nodes, a low penalty for cover nodes and a higher penalty for non-cover nodes. See docs about penalties.
However you would want to use tags to apply them (see docs about tags). Cover nodes get the tag “Cover” (or something") and non-cover nodes get the tag “OpenArea”. You could then specify on the seeker, how much penalty to apply for those tags, you would want to use a high penalty for OpenArea tag and a penalty of 0 for cover tags.

A navgraph is however not optimal for this (unless you build it yourself or make modifications to the generated one) since the nodes probably wont well represent only a “Cover” area or an “OpenArea” area, but likely be a mix of both.
A point graph would probably work, I don’t know. But if you have static environments, manually building/altering the navmesh would be the best option I think. Altering a navmesh can be done in any 3D modeling application (e.g Blender).

When you only want the shortest path, you would simply make a script set the penalty for the “OpenArea” tag to be zero as well.