> ## Documentation Index
> Fetch the complete documentation index at: https://docs.xpander.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Skills on an agent

> Give your agents the skills your workflow needs.

A skill is one outside system an agent can act in, with each of that system's operations available as an action. Admins decide which skills exist for the organization, and who sees each, in **Platform > Agents > Skills registry**; you add them to an agent with **Add skill**. An item an admin set to **Auto-injected** or **On by default** is already on every agent in its audience; an **Optional** one waits for you to add it; an **Off** one is on no agent.

![Agent settings Skills row with Core Skills, GitHub Search Manager and Google News expanded: Authentication Secret, Execution policy Mock or Live, Ask for approval switches, and the News Retrieval action group](https://mintcdn.com/xpanderai-099931d1/H-WYYv839kjeMNSx/images/image-48.png?fit=max\&auto=format\&n=H-WYYv839kjeMNSx\&q=85\&s=a2b626384d2dfd42d5ba3da2e014cc92)

A skill on an agent: its connection, Mock or Live data, approval switches and its actions by group. Shown with sample data.

## Add a skill

Use **Add skill** in the agent's settings to give an agent a skill. The panel filters by kind: `Connectors` for both API and MCP connectors, and `Skills` for Markdown skill bundles together with the built-in skills the runtime ships with (marked with a **Powered by Xpander** badge).

> [!NOTE]
>
> For the full catalog of skills you can add within xpander, see the [skills catalog](https://docs.xpander.ai/connectors.md).

### Skill visibility

| Visibility | Who can see and attach it |
| - | - |
| `org` | Everyone in the organization. The skill is in the org catalog and any agent can attach it. |
| `restricted` | An allowlist of agents. Only the agents on the list can hold the skill. |
| `agent` | One owner agent. The skill exists for that agent alone. |

**Shared with** on a skill offers **Private**, **People** (named people and groups), **One agent** and **Organization**. The `restricted` and `agent` scopes are set when a skill is published from a repository with `xpander.yaml`; Xpander Chat shows such a skill to the agents on its allowlist and refuses to attach it to any other.

As a builder you see the org catalog plus the skills scoped to agents you manage. A skill scoped to agents you do not manage stays out of your catalog. Attaching a skill to an agent that is not on its allowlist is refused, so share it with that agent under **Shared with** (**One agent**) first, or ask the skill's creator or an org admin to do it.

Who may change a scope, how the runtime enforces it, and which audit events it writes are on [Skill visibility](https://docs.xpander.ai/resources/security-compliance/access-control.md#skill-visibility). To publish skills from a repository with their scope set in `xpander.yaml`, see [Publish Skills](https://docs.xpander.ai/api-reference/v1/skills/publish.md).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.