How to combine Graph & SharePoint permission consent into a single MSAL dialog on first use

The Problem

The reason for writing this post and creating the accompanying simplified proof of concept github repo was that I was struggling to find a documented way to achieve getting the end user to consent to both Graph and SharePoint permission scopes without being presented with two consent prompts. The reason I wanted this is that I have an application that needs to make both Graph API and SharePoint REST calls before it is of any use and it is a really bad experience for the user to have to see a prompt for Graph permissions followed immediately by a second prompt for SharePoint permissions.

I have solved this problem in the past using the V1 authentication model and declaring both the Graph and SharePoint permissions on the app registration and just specifying the .default scope. This was using MSAL but really just reverting back to the legacy V1 authentication model. I wanted to solve this problem using the V2 authentication model approach of dynamically asking for scopes.

Challenge 1 – MSAL AcquireTokenSilent & AcquireTokenInteractive only support one resource at a time

Let me walk you through what I’ve found from the documentation. When using MSAL and V2 permissions (where you specify your scopes in the calling code rather than declaring all permissions on the app registration) the documented approach is to follow this pattern where you first try to AcquireTokenSilent and if the user needs to login or grant consent to one or more of the requested scopes a MsalUiRequiredException is thrown, which you catch and use AcquireTokenInteractive.

            authResult = await app.AcquireTokenSilent(scopes, firstAccount)
        catch (MsalUiRequiredException ex)
            // A MsalUiRequiredException happened on AcquireTokenSilent.
            // This indicates you need to call AcquireTokenInteractive to acquire a token
            System.Diagnostics.Debug.WriteLine($"MsalUiRequiredException: {ex.Message}");

                authResult = await app.AcquireTokenInteractive(scopes)
            catch (MsalException msalex)
                ResultText.Text = $"Error Acquiring Token:{System.Environment.NewLine}{msalex}";

The above sample code was taken directly from the official Microsoft Docs.

The problem here for my use case is that AcquireTokenSilent and AcquireTokenInteractive can only be used to get an access token from a single resource (either the Graph or SharePoint) not both at the same time. If you try to make 2 calls you end up calling AcquireTokenInteractive (passing just the Graph resource scopes) which the user gets prompted to consent to, and then the second call to AcquireTokenInteractive (passing just the SharePoint resource scopes) which the user gets a second prompt for. What I did find that has been added to AcquireTokenInteractive however is that while you can only use it to retrieve a token for a single resource you can optionally provide extra scopes for different resources to ask the user to consent to. This has been documented officially here and below is a screenshot of the relevant section of that official documentation.

Brilliant this now gives us a way to initially try to acquire the Graph token but also in the same call ensure the user has consent to the SharePoint scopes. Conceptually it should be as easy as this.

var result = await app.AcquireTokenInteractive(scopesForGraph)

Challenge 2 – The SharePoint resource endpoint is tenant specific

Now we are down into the actual format of specifying the individual permission scopes. The Graph ones are pretty straight forward they all look like this ‘‘ but what information I could find around the internet for specifying SharePoint permission scopes is the permissions are formatted like this ‘https://YOUR-SHAREPOINT-TENANT/allsites.manage‘. The permission scope itself has your SharePoint tenant in it e.g. Now this may not be an issue if you know the SharePoint tenant, but in my case I’m building a multi-tenant application and I can’t possibily know who my users are going to be and what tenant they will come from.

So here’s the dilemma:

  • I can’t prompt the user for consent to SharePoint until I know their SharePoint tenant URL
  • The way to get their SharePoint tenant URL is to call the Graph
  • I can’t call the Graph until I prompt for consent

It seems we are back to the original crappy user experience that we need to prompt for consent (just for Graph permissions) so we can call the Graph to get the SharePoint tenant URL so we can then prompt again to get consent for the SharePoint permissions before we can call the SharePoint REST API.

The Solution

Ok lot of background info, but it’s all getting us closer to the code in this repo and the challenges it solves. Next I turned my attention to the old V1 authentication model of specifying the mixed permissions (Graph and SharePoint) on the app registration and then just getting the user to consent to that. Note – the idea of this is still very relevant as this is how a tenant admin gives consent on behalf of all users. The admin consent URL just points to an app registration by its id and the admin is presented with the consolidated list of permission to grant consent to. So under that model how does the app registration know what the SharePoint scope should be for a multi-tenant app? I took a closer look at an app registration in Azure portal. This is what it looks like.

Ok so that’s intriguing, what is this, it’s not the name of my tenant that I’ve created the app registration in. My best guess is that it stands for dogfood and it’s the Microsoft SharePoint dogfood tenant. Alright so maybe there’s something magical about this tenant, let’s put it to the test. Rather than worry about trying to figure out the actual SharePoint tenant for our user let’s change the first consent prompt we show (for the Graph) and pass in this weird Microsoft tenant in the .WithExtraScopeToConsent. So now we are conceptually going to do this:

var result = await app.AcquireTokenInteractive("")

Get outta town, you know what this did? It gave me the consolidated prompt, but have a look at those SharePoint permissions, it knows who the user is, and it’s dynamically switched out that Microsoft SharePoint tenant for the users tenant without us doing anything!

Alright we’ve got some crazy undocumented magic going on here, but this does appear to give us the solution to the problem from a consent perspective. I went on to discover that we do still need to do the work of finding the users real SharePoint tenant URL because we need to use the real tenant URL in the scope when we try to acquire the SharePoint access token. Remember in that dialog above we were trying to acquire just the Graph token but passing in the extra SharePoint scope just for consent.

With all that background and comments in the code I think you’ll find your way through the proof of concept pretty easily. I intentionally kept the code as simplistic as I could (no Graph SDK, no SharePoint SDK, no error handling) just raw calls so you can see the mechanics without extra code fluff.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Create a website or blog at

Up ↑

%d bloggers like this: