The global aging population demands innovative solutions for eldercare and monitoring, with social robotics emerging as a promising tool. This study addresses the design and implementation of an advanced task planner to optimize human-robot interaction in social robots, enabling personalized and non-intrusive monitoring and assistance in the home environment. To achieve this, a social robot prototype was developed specifically designed to assist, monitor, observe, and care for older adults in their homes based on the programming and adaptation of a Temi robot. The core component of this system is its central task planner, which is responsible for orchestrating and commanding the robot's actions through the dynamic execution of a sequence of tasks. Unlike static approaches, our planner allows for the definition of customizable and adaptive tasks, which are essential for effectively responding to the changing needs of older individuals. Preliminary laboratory results demonstrate the robot's capability to propose and execute different activities (e.g., medication reminders, conducting environmental checks, simple conversations to combat loneliness, among others) adapting to the user's specific requirements.