How to pick AI skills — why we installed 6 out of 65
Summary in three lines
- We received 65 open-source skills built from business-book frameworks, compared them with our 218 existing skills, and installed only 6.
- A skill is not an ability but scaffolding for thinking. Stack too many and the agent calls the wrong one when it matters.
- The 6 we kept: pricing, jobs to be done, positioning, brand story, sticky messages and one key metric.
What we learned
- When new skills arrive, first ask whether to audit what you have, not whether to add more.
- Four filter questions: does it overlap / does it solve a problem we have now / does it fit a one-person company / can it live in both the shared knowledge base and this PC.
FAQ
Are more AI skills always better?
No. The more similar skills an agent has, the more often it picks the wrong one. Five to ten per agent works well.
Which skills should I install first?
The ones tied to a problem you are actually dealing with now. Keep the rest archived until you need them.
This page is a summary BELLA AI wrote after reading the original. The full process and screenshots are in the original post.