The rise of Large Language Models has intensified the need for reliable assessments of foundational programming skills in introductory computer science courses. While frequent, low-stakes testing is an effective pedagogical strategy, its adoption is often slowed by the need for institutional infrastructure like a Computer-Based Testing Facility, which instructors may lack or find too inflexible. This experience report presents a practical, instructor-driven model for Skill Tests: weekly, 10-minute, proctored, computer-based assessments run by course staff in any campus room using student laptops. We provide a logistical blueprint refined over five offerings of a large CS1 course, detailing our strategies for question design, student scheduling, and staff management. Our findings show this model is effective: replacing a midterm with Skill Tests reduced student anxiety, with 91% of students preferring the new format. Student performance on the final exam remained consistent, indicating that knowledge synthesis was not compromised. This paper offers educators a framework for deploying frequent, low-stakes assessments without dedicated institutional resources.