As autonomous drone swarms are increasingly deployed in high-stake environments, they become prime targets for advanced cyber threats. Traditional centralized security mechanisms fail to meet the demands of decentralized systems, which require low latency, high resilience and adaptability. SHIELD (Swarm-Enabled Hierarchical Intelligent Edge Defense) introduces a multi-layered defense system tailored to these needs, integrating Edge AI, Federated Learning (FL), Generative AI (GenAI), and Intra-Cluster Peer-to-Peer (P2P) Communication. At the edge, drones autonomously detect and neutralize threats in real-time even when communication with higher layers is compromised. The fog layer coordinates swarm intelligence by securely aggregating updates and distributing models, while the cloud layer uses FL and GenAI to continuously refine the global threat detection models based on real and synthetic scenarios. In cases of prolonged disconnection from the cloud, SHIELD transitions to intra-cluster P2P collaboration, enabling localized responses and coordination among drones. This architecture empowers drones to dynamically adapt to emerging threats, providing a resilient and scalable defense system that maintains operational continuity in high-risk environments. By redefining security for autonomous systems, SHIELD enhances the reliability of drone operations in sectors like critical infrastructure and disaster response.