blog · definition
What Is an Agent Skills Registry?
An agent skills registry is a queryable catalog that AI agents search at runtime to discover and apply skills for a given task. Unlike a skill marketplace — which requires a human to manually install skills first — a registry is designed for direct agent consumption. The agent queries, the registry responds, the skill is applied. No human in the loop.
Registry vs Marketplace
| Marketplace | Registry | |
|---|---|---|
| Designed for | Human browsing | Agent queries |
| Discovery | Human searches | Agent queries at runtime |
| Installation | Required | Not required |
| Query method | UI / search | Programmatic API / MCP |
| Example | skills.sh | Invoked |
How it works
A registry exposes a search endpoint. An agent describes its current task context — what it's trying to do, what kind of skill would help. The registry returns matching skills ranked by relevance. The agent fetches the skill content and uses it to guide its behavior.
Invoked implements this as an MCP server. AI agents connect once via standard MCP configuration. Every query searches the full registry — returning skills from any contributor, any domain, any use case. Configure once; your agent discovers skills forever.