Bricklaying faces significant labor shortages, for which Mobile Construction Robots (MCRs) offer a promising solution. However, the absence of many building elements and hazards in brickwork environments, such as floor and exterior wall openings, which may be undetectable to robot sensors, poses safety challenges to the navigation system. This paper presents an automated workflow to integrate IFC-based Building Information Modeling (BIM) information into the Robot Operating System 2 (ROS 2) navigation stack (Nav2) by generating compatible 2D Occupancy Grid Maps (OGMs) and safety-zone masks. To do so, navigation-relevant geometric, semantic, and georeferencing information is extracted from the IFC model and stored in dedicated data structures. Then, detectable and undetectable obstacles are classified based on the robot's mechanical structure and sensor configuration, including the traversability threshold and the 2D LiDAR scan plane height. This results in the generation of (1) a static main map containing detectable obstacles and (2) a keep-out mask encoding undetectable hazards, both of which are rasterized automatically and integrated into Nav2, where keep-out regions are enforced as costmap filters for global path planning and local path tracking. Simulation results demonstrate that this approach enhances operational safety in brickwork construction environments by increasing the safety success rate and eliminating intrusions into the safety zones.