Product Story

Why We Built SkillRegistry

The journey from scattered skills to a home for AI workflows.

SkillRegistry hero

SkillRegistry started with a simple frustration. I kept bumping into new skills for agents, but every one of them lived somewhere else. Some were hidden in GitHub repos, others were tucked away in random gists, and a few only existed in private notes. I wanted to move faster, but every workflow started with a scavenger hunt.

When I couldn't find what I needed, I built my own skills. That meant copy-pasting, testing, and rewriting the same logic over and over. There was no consistent place to share or compare skills across teams, so every project felt like starting from zero.

The harder part was that LLMs are still learning how to build skills well. Skills are new, conventions are still evolving, and the best practices aren't obvious. Models can draft the basics, but they don't yet have the lived experience of using skills in production. That means people still need a human-curated place to learn what good looks like.

SkillRegistry is our answer. It's a shared library for real skills that you can search, compare, and plug into your workflow. Our goal is to save teams from the same fragmented journey and give this new ecosystem a real home.