A skill that reconstructs how you actually wrote something with AI, then turns the transcript into two visual one-pagers. One is a candid self-audit. The other is a transparency report you can share. It is free to install.
A flat tally of turns tells you nothing. What makes the analysis worth reading is the evaluative spine: it measures your actual process against a research-grounded bar for healthy AI collaboration, then judges it honestly, including pushing back on its own conclusions. The numbers serve the story; they are not the story.
The question it answers is the one that should worry anyone using AI to write: did I do the thinking, or did the tool do it for me? The analysis makes that visible instead of leaving it to a vague feeling.
I ran this on my own work and wrote up what the transcript showed. Read it on Substack →
The full read. Cognitive-protection scoring, the qualitative story of how the piece came together, an honest critique, and the charts. The candid self-audit you would not necessarily show anyone.
Bare bones and safe to share with an editor, a collaborator, or your readers. Turns, division of labor, tools used, and one concrete rigor proof-point. Omits the introspection.
PERISCOPE .skill file above.The scoring rubric is editable, so you can hold your own work to your own bar. The thumbnail above is a real output from this skill, an analysis of a Substack piece I wrote.