The vast majroity of tasks work without issue. But there's a lot of edge cases where uutil fails. When those are found, new tests are added to track them and prevent future regressions.

And new test cases are constantly being added. And with those new cases, the pass rate drops.