We propose a novel approach to agent placement in MANETs based on node vulnerability. A cost function is introduced to quantify the vulnerability of each node, reflecting its role in maintaining overall network connectivity. Building on this, we model the agent placement problem as a non-cooperative game, where players seek Nash equilibrium strategies for locating a new agent in an unoccupied node. In disaster-recovery scenarios, these agents may represent rapidly deployable UAVs (drones) acting as mobile relay nodes to extend and stabilize network coverage. To validate the theoretical results, a real-world network topology from New Orleans during the post-disaster recovery efforts following Hurricane Katrina in 2005 was used. Simulation results demonstrate that positioning agents according to the proposed algorithm leads to a measurable improvement in key network performance metrics.