What you get
It tells you what it assumed, instead of folding the assumption into the conclusion.
31 Data-Analytics Skills, Scripts Included
Added 2026-08-06Organised into six groups — data quality, documentation, analysis, visual storytelling, stakeholder communication, process — and what ships in the repo is runnable Python, not just prompts.
The scripts are real: null counting, duplicate finding, referential-integrity checks, A/B test analysis and SaaS metric calculation each have one. The interaction is fixed in four steps: ask for the minimum usable context, run a structured process, state assumptions, output to a template. The author is clear that it degrades gracefully — given incomplete context it says what it is assuming and continues, so conclusion quality tracks how much you fed it. Customisation is dropping reference files into any skill.