> For the complete documentation index, see [llms.txt](https://docs.ox.security/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.ox.security/vibesec/agent-ai-bom/agent-ai-bom-mcp-servers.md).

# Agent AI BOM: MCP Servers

On the MCP Servers tab of the [Agent AI BOM](/vibesec/agent-ai-bom.md) page, you can track which MCP servers agents connect to, which tools those servers expose, how frequently they are used, and the developers who use them.

Reviewing this list lets you confirm that only approved MCP servers are in use, identify newly added or unexpected MCP servers, and determine which developers are responsible for each connection so you can follow up where appropriate.

Summary tiles at the top of the tab show totals for MCP servers, active users, and calls within the current scope.

<figure><img src="/files/UOBx2TwpwxZvR1I2LHZS" alt="" width="563"><figcaption></figcaption></figure>

## Table columns

<table><thead><tr><th width="170.66668701171875">Column</th><th>Description</th></tr></thead><tbody><tr><td>Name</td><td>The name of the MCP server.</td></tr><tr><td>Type</td><td>The connection type of the MCP server. Common values are <code>http</code> and <code>plugin</code>.</td></tr><tr><td>Identifier</td><td>The unique value OX uses to identify the MCP server. This is typically the MCP endpoint URL, a plugin identifier, or another unique ID.</td></tr><tr><td>Total Users</td><td>The number of distinct developers in your organization who have called this MCP server.</td></tr><tr><td>Total Calls</td><td>The total number of tool calls made through this MCP server.</td></tr><tr><td>First Seen</td><td>When OX first recorded a call to this MCP server.</td></tr><tr><td>Last Activity</td><td>When the most recent call was recorded. Use this to spot MCP servers that are no longer in use, or that were just added.</td></tr></tbody></table>

## Filters

<table><thead><tr><th width="199">Filter</th><th>Description</th></tr></thead><tbody><tr><td>MCP Name</td><td>Limits the list to selected MCP servers by name.</td></tr><tr><td>MCP Identifier</td><td>Limits the list by the MCP server's identifier, which is typically the MCP endpoint URL or a vendor-supplied identifier string.</td></tr><tr><td>User</td><td>Shows only MCP servers used by selected developers.</td></tr><tr><td>Tool Name</td><td>Shows only MCP servers that expose the selected tool. Useful for finding every MCP server that exposes a specific capability, such as a write tool.</td></tr></tbody></table>

## Reviewing MCP server activity

The **Explore** tab provides an overview of how an MCP server is used across your organization. Use this view to understand adoption, identify the most frequently used tools, and determine which developers rely on the MCP server most heavily.

<figure><img src="/files/VIPvH0IsdbzkrSOoBpzY" alt="" width="563"><figcaption></figcaption></figure>

<table><thead><tr><th width="159.83331298828125">Section</th><th>Description</th></tr></thead><tbody><tr><td>Summary metrics</td><td><p>Provides a quick overview of MCP server activity within the current scope, as follows:</p><p>• <strong>Total Calls</strong>: The total number of tool invocations made through the MCP server.</p><p>• <strong>Total Users</strong>: The number of developers who used the MCP server.</p></td></tr><tr><td>Most Used Tools</td><td><p>Shows the most frequently used tools exposed by the MCP server.</p><p>Using this section, you can:<br>• Quickly identify the tools developers use most often.<br>• Understand the primary use cases of the MCP server.<br>• Spot heavily used tools that may warrant closer review.</p><p>To view all tools exposed by the MCP server, select <strong>View All</strong>.</p></td></tr><tr><td>Most Active Users</td><td><p>Shows the developers who generated the most MCP activity.</p><p>Using this section, you can:</p><p>• Identify which developers rely most heavily on the server.</p><p>• Investigate unusual usage patterns.</p><p>To view all users who accessed the server, select <strong>View All</strong>.</p></td></tr></tbody></table>

## Blocking tools

You can prevent AI coding agents from using tools whose access scope is not appropriate for your organization. For example, you might block tools that can access sensitive systems, perform administrative actions, or expose data that developers should not be able to retrieve through AI agents.

When you block a tool, OX intercepts and denies future attempts to invoke that tool. The block can apply to a single tool, multiple tools, or all tools exposed by a specific MCP server.

From the developer's perspective, the tool becomes unavailable. If the agent attempts to invoke a blocked tool, the request is denied, and the action is not performed.

Each block is recorded on the [Prevented Risks](/vibesec/vibesec-prevented-risks-new.md) page, where you can review which tools were blocked, for which developers, and when.

**To block a tool:**

1. Go to **Agent AI BOM** > **MCP Servers**.
2. Select the required MCP server.
3. Open the **Tools** tab.

<figure><img src="/files/xbJsxczawv7bmgc79SEI" alt="" width="563"><figcaption></figcaption></figure>

4. Locate the tool you want to restrict.
5. To block the tool, turn off its toggle.

After a tool is blocked, OX records each blocked invocation attempt on the **Prevented Risks** page. You can review which tool was blocked, which developer attempted to use it, and when the attempt occurred.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.ox.security/vibesec/agent-ai-bom/agent-ai-bom-mcp-servers.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.