A program like this is not full automation, but it greatly reduces the cognitive overhead of doing this task. Now when I suspect a sensor is bad, I can just pull out the script without having to get back up to speed on how to orchestrate the component programs and interpret the results. It's like having a well-designed jig for a woodworking task, hanging on the wall for later use.
Writing tiny programs like this is really satisfying because the scope is small enough that I can carefully consider all decisions. I spent a lot of time reading through guides and code samples, considering dependencies vs pure bash, understanding shellcheck warnings, and making the code conform to a style guide. All not strictly necessary to achieve the goal, but I learned a lot and take satisfaction in a small, but refined result.