> For the complete documentation index, see [llms.txt](https://docs.devolutions.net/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.devolutions.net/powershell-universal/intelligence/built-in-mcp-tools.md).

# Built In MCP Tools

PowerShell Universal includes several built-in MCP tools that help coding agents inspect PSU resources and learn the local command surface. These are available from the `/api/v1/mcp` endpoint.

These built-in tools are separate from your custom [AI Tools](/powershell-universal/intelligence/ai-tools.md). In practice, they are most useful when an agent needs to inspect scripts, endpoints, apps, or cmdlet help before making a change.

To access them remotely, connect an MCP client to `/api/v1/mcp` and authenticate with a PSU app token. The built-in tools require `Administrator` or `Operator` access.

## Typical Flow

Most coding sessions follow a pattern like this:

1. Call `list_resource_types` to see what kinds of PSU resources are available.
2. Call `list_resources` to find the specific script, endpoint, app, or workflow you need.
3. Call `list_commands` or `list_app_commands` to discover the cmdlets available on that PSU server.
4. Call `command_help` for focused help on the cmdlet you plan to use.
5. Call `reload_resource` after changing configuration so PSU reloads the resource.

## list\_resource\_types

Returns the resource types available on the connected PSU instance, such as scripts, endpoints, apps, and workflows.

Use this first rather than guessing the available type names.

Example:

```
list_resource_types
```

## list\_resources

Lists resources of a given type and returns them as JSON.

Parameters:

* `type` - The resource type to inspect. Use a value returned by `list_resource_types`.
* `filter` - A wildcard filter that matches resource names or descriptions.

Example:

```json
{
	"type": "<resource type from list_resource_types>",
	"filter": "deploy*"
}
```

## list\_commands

Searches for commands in a PowerShell module installed on the PSU server.

Parameters:

* `module` - The module name.
* `name` - A command name filter. Wildcards are supported.

This is usually followed by `command_help`.

Example:

```json
{
	"module": "Universal",
	"name": "*PSUAi*"
}
```

## command\_help

Returns help information for a command available on the PSU server.

Parameters:

* `module` - The module name.
* `command` - The command to inspect.
* `parameter` - Optional parameter name or wildcard pattern.
* `full` - Include the full help output.
* `examples` - Include examples.

Start with targeted help instead of always requesting the full help text.

Example:

```json
{
	"module": "Universal",
	"command": "Invoke-PSUAiAgent",
	"parameter": "Prompt",
	"full": false,
	"examples": true
}
```

## list\_app\_commands

Returns the app-specific cmdlets used to build PSU apps, especially `UD` component and interactivity commands.

Parameter:

* `filter` - A wildcard filter for the command name.

Example:

```json
{
	"filter": "New-UD*"
}
```

## reload\_resource

Reloads a PSU resource type after files or configuration have changed.

Parameter:

* `resourceType` - The resource type to reload.

Example:

```json
{
	"resourceType": "<resource type from list_resource_types>"
}
```

This is especially useful when an agent edits configuration through a local checkout or remote file system and then needs PSU to pick up the change.


---

# 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:

```
GET https://docs.devolutions.net/powershell-universal/intelligence/built-in-mcp-tools.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
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.