While this is true and solid advice, in the greater context of the entire article, it perpetuates an implicit assumption that all of accessibility or even more narrowly, all of WCAG is programmatically testable. It isn’t. Only roughly half of the 61 SC of 2.0 can be audited by a machine. The rest requires a human evaluator.

I really love the granularity of the tools evaluation. Perhaps a table view by features could be another way to provide your Tl;DR conclusion list.