Robotic-assisted surgery has significantly enhanced precision, dexterity, and ergonomics in minimally invasive procedures. However, surgical suturing remains one of the most challenging tasks to automate due to its requirement for fine manipulation, bimanual coordination, and robustness to geometric and visual variability. Recently, Reinforcement Learning has emerged as a promising approach for autonomous surgical skill acquisition. Progress is often limited by the lack of high-fidelity simulation environments that accurately reproduce the kinematics, constraints, and control interfaces of real surgical systems. In particular, existing simulators for the da Vinci Research Kit (dVRK) frequently rely on simplified models and lack seamless integration with teleoperation and learning pipelines, increasing the sim-to-real gap and limiting experimental reproducibility.