Methodology

How this catalogue is built

This page exists so that a reader can judge the catalogue instead of trusting it. It describes what was read, what was tested, and what a desk review cannot prove.

How skills are selected

Every skill here is published openly by a named author, is installable today, and solves a problem an agent user meets in normal work. Skills that exist only as a screenshot, a private folder or an announcement are not listed.

What was actually read

Each entry was written after reading the skill's own repository or directory listing, including its stated license and its own description of scope. Where the publisher does not state something, the entry says so rather than guessing.

What the order means

The order inside a category reflects how broadly useful the skill is to a working reader, not popularity. Star counts and install counts move daily, cannot be re-verified for thirty entries on every build, and are therefore not published here at all.

What this does not prove

A desk review is not a security audit, a benchmark or a guarantee of maintenance. Read the source before you install anything into an agent that can run commands on your machine.