<p>Health data sharing is essential for enhancing healthcare services, but unauthorized access remains a significant issue, especially for time-series data generated by wearable monitoring sensors. Traditional methods such as encryption and access control often fail to mitigate access anonymity during real-time data transmission across IoT–cloud platforms. This study presents ATTUNE-SHARE, a novel framework based on the Migration Agent Learning Algorithm (AS-MALA), designed to improve data security by ensuring the secrecy of time-series physiological signals including heart rate, pulse, and oxygen saturation. The framework employs migration-based learning to dynamically adjust secrecy weights assigned to agents, enabling secure agent movement toward high-weighted nodes to prevent unauthorized access. The secrecy rate is computed using these adaptive migration agents with the objective of maximizing confidentiality. To strengthen the cryptographic protection, the framework integrates AES-256, a NIST-standardized and publicly verified encryption method that ensures strong confidentiality and integrity during IoT–cloud communication. AES-256 provides robust defense against brute-force and statistical attacks while maintaining low computational overhead, making it suitable for resource-constrained healthcare devices. Experimental results demonstrate that AS-MALA improves anonymity detection by 14.89% and enhances secrecy rate by 11.95% compared to existing baselines. The findings validate the capability of ATTUNE-SHARE to secure real-time healthcare data sharing while preserving privacy in dynamic and open IoT–cloud environments.</p>