My writing linter can be defeated by writing more. I measured how much more.

I ran my writing checker on a published post twice: first on the entire file, then on just the body with the front matter removed. The scores were the same, but the verdict crossed a publish gate only on the body. Stripping the front matter contributed words and dashes to the denominator, allowing the post to pass the gate. The tool measures two factors: set-based word-level tests and density-based em-dash and bold tests.

The density tests can be driven down by adding non-objective text, while the set-based tests are unaffected by repetition. The checker combines these two factors into a single score, but they should be scored separately. The front matter, titles, and descriptions are metadata not to be averaged with the main body. Splitting the score into two scalar values, one for each family, would prevent padding from inflating the denominator and provide a more accurate assessment.

The current system produces a misleading score that is trusted for the wrong reason, leading to a flawed publishing gate.