The growing confidence on IoT devices such as the Raspberry Pi 3 Model B requires energy-efficient CPU management to balance the performance and the longevity of the CPU life. The default on-demand governor, widely used in these devices, suffers from frequent and abrupt frequency changes, leading to higher power consumption, increased thermal stress, and long-term CPU degradation. These rapid alternation in CPU frequency accelerate electro-migration, a phenomenon where repeated voltage changes degrade transistor pathways, ultimately reducing processor lifespan. This paper introduces a lightweight a task-based CPU frequency scaling framework that overcome these issues by reducing frequency jumps by 90%, thus stabilizing performance and improving energy efficiency. Experimental results show that, unlike ondemand, which shows extreme frequency variation, the proposed framework performs a more stable frequency, significantly reducing CPU stress. Moreover, it enhance execution time and lowers average CPU temperature, showing a 6–8% reduction in temperature compared to the state-of-the-art. Unlike machine learning-based CPU governors, which introduce computational overhead, this proposed framework confirms real-time, low-cost dynamic frequency adjustments through a daemon-driven architecture. By offering a developer-friendly API for workload-specific tuning, this framework presents a practice, low-overhead alternative for efficient power management in constrained IoT and edge computing environments.