Posts

Based on one of my older posts about PowerShell script signing with Azure DevOps I recently implemented a PowerShell script signing workflow with GitHub actions I wanted to share with the community. Prerequisites # For this post the following prerequisites are required: Code signing certificate in PFX format GitHub account Add variables to GitHub actions # Because GitHub variables can only be of string content we need to get the contents of the pfx file as base64 encoded string: $pfxCertFilePath = "~\Downloads\CodeSigningCertificate.pfx" $pfxContent = Get-Content $pfxCertFilePath -Encoding Byte [System.Convert]::ToBase64String($pfxContent) | Set-Clipboard GitHub action variables # Add two variables as actions secrets: BASE64_PFX: Base 64 encoded string of the PFX (automatically copied into your clipboard with the above commands) PFX_PASSWORD: Password for the private key of the pfx file GitHub action workflow # Place the following workflow file within your git repository: .github/workflows/SignPowerShell.yaml whereas the name of the YAML file can be freely chosen.

The Send-MailMessage cmdlet has been around for a couple of years and is mostly used to send email messages from PowerShell. But with the deprecation and security flaws of legacy authentication it’s time for a better option which actually supports modern authentication. For this purpose we can use the Microsoft Graph API and the Microsoft Graph PowerShell SDK. The best thing is that this solution works without any service account and does not need any exclusions from conditional access. Microsoft Graph resource # To send a mail we simply specify the user account from which we want to send the email: POST: https://graph.microsoft.com/v1.0/users/[email protected]/sendMail Create an app registration # Simply create a new app registration with the Mail.Send permissions and use a certificate for the authentication. We need to take additional steps to limit the permissions of the app registration. Otherwise the app can send mails on behalf of any user in your tenant. To limit the permissions we leverage exchange application access policies. Connect to Exchange Online with the ExchangeOnlineManagement PowerShell module Connect-ExchangeOnline

While fine-tuning and adjusting applocker policies for co-managed Windows 10 clients I got really annoyed by special characters commonly used in the German/Swiss language. The Intune portal seemed to use different encoding and didn’t allow me to just copy/paste the currently deployed policy and extend it with a new rule. I needed to request the original file that was uploaded to the tenant in order to adjust the rule. Instead of just accepting this I decided that it is time for an easier approach which I will share with you. The actual issue # After uploading the XML with the applocker policies (in my case EXE rules), special characters like ‘ö’ or ‘ü’ have a weird encoding displayed in the portal. The next person that wants to edit the policy needs to take care to fix the unrecognized characters otherwise the publisher rule won’t work anymore. Top: Portal view of the special characters, bottom: original file. Fixing things # Save file with UTF-8 encoding # First, make sure that you saved an uploaded the file with UTF-8 encoding as this introduces support for most character sets. You can do this by selecting the encoding box in the right bottom of Visual Studio Code:

When working with the Microsoft Graph API or introducing the API to colleagues I often get asked about the steps required to obtain an access token for the API with PowerShell. Out in the wild, I’ve spotted many different ways and lots of implementations still relying on the ADAL (Active Directory Authentication Library) despite the fact that this client library is superseded by MSAL (Microsoft Authentication Library). So let’s talk about acquiring access token “in stile” with the most simple method available. Why do we need an access token? # When talking about the Microsoft Graph API an access token fulfills two roles, first: prove authentication (proof of identity) second prove authorization (permissions). Each request needs to submit a request-header that contains the access token. For an API it’s crucial to validate the authentication and authorization for every request. Otherwise, requests could be made to resources the actor has no access to. What’s inside the access token # You did probably stumble over the terms “bearer authentication” or “bearer token” these describe a mechanism within the OAuth 2.0 Authorization framework to authenticate requests with access tokens. Tokens are issued by the authorization server (Azure AD) and contain a server-generated string in the format of a JSON Web Token (JWT) with the following information (the list is not exhaustive and truncated to only contain the most interesting parts):

While doing some Android Enterprise enrollment tests for corporate-owned devices with work profiles I stumbled over the following issue after signing-in with the work account: “Page not found”. The solution is fairly simple, just double check that your user does not have the device enrollment manager role assigned, which can be found under the device enrollment pane: The docs tell us: If you’re enrolling Android Enterprise personally-owned work profile or corporate-owned work profile devices by using a DEM account, there is a limit of 10 devices that can be enrolled per account. Microsoft Docs In my case, I wasn’t exceeding that limit because it was the first enrollment with that account so I’m not sure if the docs are accurate. Updated 21.12.2020: I already opened an issue on GitHub about the doc contents. @nicolonsky Good question. Using a DEM account isn’t available to enroll COPE devices. We fixed the article. The changes should be live later today. Thanks for bringing this to our attention. Hope this helps.

I like to have a linux machine for some lab stuff which I can access from multiple machines prefereably over SSH. Because Windows 10 ships with an integrated SSH client and Windows Terminal looks just awesome I wanted to use Windows Terminal to access my linux machine running on Azure over SSH. Today I’d like to show you my setup. Generate a Key Pair # ssh-keygen Generating public/private rsa key pair. Enter file in which to save the key (C:\Users\NicolaSuter/.ssh/id_rsa): Created directory 'C:\Users\NicolaSuter/.ssh'. Enter passphrase (empty for no passphrase): Enter same passphrase again: Your identification has been saved in C:\Users\NicolaSuter/.ssh/id_rsa. Your public key has been saved in C:\Users\NicolaSuter/.ssh/id_rsa.pub. Add SSH config file # C:\Users\%USERNAME%\.ssh\config Host horus Hostname horus.nicolonsky.ch Port 22 User azureuser IdentityFile ~/.ssh/id_rsa Test the SSH connection # ssh horus Windows Terminal Configuration # Create a new GUID with Powershell and the command: (New-Guid).Guid { "guid": "{67d39a21-70ff-4bdd-af09-fd68b29c1716}", "name": "Horus", "commandline": "ssh horus", "hidden": false, "icon": "https://image.flaticon.com/icons/png/512/119/119423.png" }, Sync Windows Terminal Settings with OneDrive # We can setup a symlink to use the Windows Terminal Config stored in OneDrive:

When involved in new projects I often find a bunch of old profiles in the Microsoft Endpoint Management Console. Before going ahead with a new implementation it’s the best time to clean-up all the leftovers from past ramblings. How to identify stale profiles # If one or multiple statements are met for a profile it is very likely to be a stale profile: No assignments, assignments to a group without members “Test” included within the profile name or description Last modified points back in time for more than a year No devices reported success/failure status for the given profile type What to do with stale profiles # So let’s be brave and delete them. But Intune doesn’t offer any [CTRL] + [Z] or recycle bin possibilities so we might want to have some kind of archive, just in case? Let’s agree that we: Check the points from the list above Ask our colleagues if they know something about the profiles and their usage Take a backup deleting them afterward is a reasonable action which is probably beneficial for everyone.

Recently I got a DM on Twitter with a question about how to export and import Endpoint Security profiles with Microsoft Graph. Besides a technical answer which might be of interest for you, I’d like to show you the workflow I used to give a proper reply. Original question: Hi @nicolonsky, I was advised on the MS Elite Partner focus groups team (MEM Automation) to reach out to you regarding my question about export/import policies from Endpoint Security in Intune. I’ve been able to export the Disk Encryption policy (via graph explorer), but haven’t been able to find the correct format to use to upload/import it. I was hoping that you would be able to advise on how to go about achieving this. Workflow # Discover request URL’s and payload # To discover the request URLs and payloads I used the methodology I explained in the this post a while ago. Basically, I tracked the network activity and used a filter to only include requests made to the Graph API while doing the following activities:

I recently bought a Surface Pro 7 with an Intel Core i7 and 256 GB SSD and it was a quite good deal. I’m using it primarily for my studies in computer sciences which involves lots of development with Java, Python, C, Linux and so on. Furthermore, I’m using it to write blog posts for this blog. One thing which annoyed me since day 1 was the fan going absolutely crazy when plugged to AC power. Having only OneNote open and a couple of browser tabs and maybe teams the noise was way too loud and even heard by other participants in meetings. On Reddit I found a hint to a setting which indeed solved my solution. We need to adjust the “max. processor power state” setting which is not visible by default on surface devices. To show the setting add this reg key with PowerShell: Set-ItemProperty -Path "HKLM:\SYSTEM\CurrentControlSet\Control\Power\PowerSettings\54533251-82be-4824-96c1-47b60b740d00\bc5038f7-23e0-4960-96da-33abaf5935ec" -Name Attributes -Value 2 Afterwards, we can change the “max. processor power state” setting within the power settings in the control panel: Hope this helps you as well to enjoy your Surface Pro 7 with a much quieter fan. 🐱‍💻

Too lazy to sign your PowerShell scripts? Yes of course it provides security benefits but performing the steps manually can be easily forgotten and re-signing needs to happen after every script change. Because I like CI/CD topics and have not found a solution on the internet I decided to build a solution based on Azure capabilities. Furthermore, I wanted a solution which does not require to hand out the code signing certificate to the respective script author which can be useful if you have a bunch of people writing PowerShell scripts. From a personal perspective, I would also recommend signing scripts you hand over to customers to ensure the integrity of the scripts because as soon as the script gets changed the signature is invalid. You can find more general recommendations about script signing in the PowerShell docs. Solution overview # The key vault will store the code signing certificate with an access policy that allows access from the Azure DevOps pipeline. The pipeline consists of the following steps: Import code signing certificate The certificate is supplied as secret in a variable group which is linked to the key vault Access to the key vault is granted with a service connection (service principal) Sign PowerShell scripts which contain a “magic token” All *.ps1 within the attached repository will be enumerated To control which scripts will be signed only those with the “magic token” get processed The “magic token” gets removed before signing the script Publish signed PowerShell scripts as pipeline artifacts To make them available for download