One job, done remarkably well

Every Skill Tree skill has one job. Not two, not "a few related things" — one. That's a design constraint, and we hold it deliberately.

Why narrow beats broad

A general-purpose system prompt tries to make the AI good at everything. It ends up average at most things and excellent at none. A narrow prompt can optimize for the specific output format, the right tone, the correct domain vocabulary, and the edge cases that actually come up in practice — because there's only one thing to get right.

Compare "help me write better" to "write a cold outreach email for a B2B SaaS company, second-touch in a sequence, assume the prospect opened the first email but didn't reply." The second prompt produces a useful result. The first produces a question asking you to clarify.

The cost of generality

Every time you ask Claude or ChatGPT to do something without context, you're paying a tax: the AI has to infer your domain, your format preferences, your audience, your constraints. A well-written skill prepays that tax once. Every use after that is at full speed.

What this means for the catalog

We ship a lot of skills — currently 50+, growing weekly. Some people ask why we don't just build one "super skill" that handles everything. Because that's not how expertise works. A surgeon and a GP are both doctors. You don't want the surgeon's precision from someone optimized for breadth.

Each skill in the catalog is the surgeon. If you need ten surgeons, you buy ten — or you grab the all-access plan and get all of them.

The right skill for the right job

Browse the catalog and you'll notice the categories: Writing, Development, Marketing, Healthcare, Legal, Business, and more. Within each category, every skill targets a specific task. That specificity is the product. It's what makes the output feel like it was written by someone who actually knows the field — not just a language model making its best guess.