Wireless networks, like wireless sensor networks (WSNs) and mobile ad hoc networks (MANETs), are an essential part of the modern communication systems and serve purposes, such as IoT, smart cities, and critical infrastructure monitors. Nevertheless, they are highly susceptible to security risks like blackhole, wormhole and Sybil attacks due to their decentralized and dynamic nature, and have energy efficiency and scalability problems. The current routing protocols are mainly concerned with performance or security improvement on a case-by-case basis which results in efficiency versus robustness trade-off. The present paper discusses this issue by suggesting a safe routing model that is founded on a trust-sensitive Connected Dominating Set (CDS). The aim of the research is to come up with a routing backbone that will reduce the communication overhead and at the same time increase resilience against malicious nodes. It uses a methodology that models the wireless network as a graph and defines the construction of CDS as an optimization problem with node trust, link reliability, and cost of communication. Secure CDO is built by a multi stage algorithm that includes node selection, pruning and connectivity enforcement that is followed by secure route discovery and lightweight encryption and authentication. Simulation-based performance assessment of the proposed method is done under different attack conditions and performance of the proposed method is compared with the traditional protocols like AODV, DSR and the current secure routing protocols. The findings indicate that the proposed framework attains a superior Packet Delivery Ratio (up to 98% in normal circumstances and much higher in attacks), a lower routing overhead (by an approximation of 30-40%), and a lower energy consumption with respect to the different sizes of the network.