Recent research in Defense Mechanisms Against Black-Hole and Gray-Hole Attacks in Mobile Ad Hoc Networks (MANETs) focuses on developing intelligent and hybrid techniques to detect and isolate malicious nodes that disrupt packet forwarding and degrade network performance. Modern approaches combine anomaly detection, trust evaluation, and cryptographic validation within routing protocols like AODV and DSR to identify suspicious nodes based on packet drop ratios, route reply behavior, and node cooperation levels. Hybrid schemes integrating machine learning models (SVM, CNN, Decision Tree) with behavioral analysis have shown high detection accuracy for both black-hole and gray-hole attacks. Cross-layer defense frameworks utilize data from the physical, MAC, and network layers to enhance detection precision and minimize false positives. Recent advancements emphasize lightweight, energy-efficient, and real-time detection mechanisms, ensuring secure and reliable communication in dynamic MANET environments prone to internal and external attacks.