One skill, for the last moment before a page goes live, when the cheapest possible time to catch a staging noindex is still available.
Most SEO skills in circulation are between 300 and 1,500 words of advice that stopped being true some time ago. If one tells you to hold keyword density between one and two percent, to sprinkle in LSI keywords, to optimise First Input Delay, or to add HowTo markup, it was written from folklore rather than from anything current, and the rest of its advice deserves the same suspicion.
We wrote two skills for this category and tested both. One is published and one is not, and the one that is not is the better story.
The skill that did not ship was about deciding what to publish. The idea was to read the results page before choosing a format, because when every result for a query is a free tool, publishing a two thousand word guide produces something genuinely good that will never rank. We tested it against real search results pulled live. It produced a real improvement: worst case 16 of 19 became 18, at p=0.0317, which is statistically significant on five runs an arm.
It is still not here. Before running anything we fixed the size of the effect a skill has to produce, and this one came in under it: the median moved by 1 where we required 2. Moving that line afterwards, having seen the number, would make every other figure on this site worthless, so the line did not move. A skill that was significant and still cut is the clearest illustration we have of why the threshold is set in advance.
The skill that shipped was the pre-publish audit, and it earns its place by noticing things nobody asked about: the second H1, the missing canonical, the Open Graph image declared as a relative path, the meta description that is really the site tagline. It is also the only skill here that makes the output measurably noisier, and its page leads with that rather than burying it.
Each skill was given a realistic task, then the identical task was run again with the skill removed. The two outputs were compared against a fixture with a known set of planted defects, so the difference is countable rather than a matter of opinion. If a skill did not make the output materially better, it is not on this page.
Each skill page carries the specific result: what the run without the skill found, what the run with it found, and where it still falls short. Back to all skills.