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

# MCP and Skills

> Add reusable tools and context to your Agents.

MCP servers and Skills expand what your Agents can do and know. Both are managed in the Config Center, and both are enabled **per Agent**.

## MCP servers

MCP (Model Context Protocol) servers give Agents extra tools. Vibex supports three transports:

| Transport | When to use it                                                                     |
| --------- | ---------------------------------------------------------------------------------- |
| **stdio** | A local process. Provide the launch command, arguments, and environment variables. |
| **HTTP**  | A remote service. Provide the URL and headers.                                     |
| **SSE**   | A remote service over Server-Sent Events. Provide the URL and headers.             |

### Add an MCP server

1. Open **Config Center → MCP** and click **New MCP server**.
2. Enter a **Name** and choose the transport.
3. Depending on the transport, fill in **Command** / **Arguments** / **Environment**, or **URL** / **Headers**.
4. Choose the scope: **User** or **Workspace**.
5. Save, then click **Validate** on the resource card.
6. In the **Agent enablement** area below the card, switch on each Agent that should use it.

<Info>
  Validation is **offline**: it only checks required fields and the URL format (HTTP/SSE must start with `http://` or `https://`). It starts no processes and makes no network requests, so passing validation does not mean the server is actually reachable.
</Info>

<Warning>
  MCP tools can read files, access the network, or modify external systems. Confirm the server's origin and permissions before you attach it to an Agent.
</Warning>

## Skills

A Skill is a `SKILL.md` file plus the supporting resources it references. It gives an Agent reusable domain knowledge or a repeatable procedure. Sources can be:

| Source         | Description                |
| -------------- | -------------------------- |
| `manual`       | Written directly in Vibex. |
| `local_folder` | A local folder.            |
| `git_repo`     | A Git repository.          |
| `marketplace`  | A marketplace source.      |

### Add a Skill

1. Open **Config Center → Skills** and click **New Skill** or **Import Existing Skills**.
2. Fill in **Name**, **Description**, and **Instructions** (the body of `SKILL.md`).
3. Save and **Validate**.
4. In **Agent enablement**, switch on each Agent that needs it.

Once configured, type `$` in the Composer to find and insert an enabled Skill.

## Export to Agent-native configuration

Both MCP and Skills can be written into an Agent's own configuration files, so the Agent's other entry points (such as its external CLI) use the same configuration.

On a resource card, use **Preview export** → **Apply** → **Rollback**:

1. **Preview export** lists the files that will be changed or created.
2. Review the list, then **Apply**.
3. If something goes wrong, restore from the same place with **Rollback**.

The write target depends on the Agent—for example Codex's `~/.codex/config.toml`, Claude's `~/.claude/settings.json`, and each Agent's Skills directory.

<Warning>
  **Always review the preview before you Apply.** By default Vibex does not rewrite an Agent's native home configuration; it writes only when you explicitly apply.
</Warning>

## Scope and safety

Use the smallest scope so you do not expose project-specific tools to every project. MCP servers currently offer only the **User** and **Workspace** scopes.

Strip tokens, private keys, environment variable values, prompts, and file contents from any logs or diagnostics.