Security & Compliance
Security operations for Microsoft 365, checked regularly.
Maester checks your Microsoft 365 tenant regularly with open, traceable tests. We pass on every deviation, follow up Defender incidents and report monthly.
For IT teams with an agreed security baseline in Microsoft 365 who want to spot deviations after changes early.
- Maester
- Maester Cloud
- Pester
- Microsoft Defender XDR
- Microsoft Secure Score
- Entra ID
- GitHub Actions
- Azure DevOps

A test fails on Monday
Monday morning: the weekly Maester run reports in your Teams channel that a conditional access policy has excluded an entire group since a change on Thursday. The person in charge opens the report, sees in the failed test which setting was expected, finds in the audit log who made the change and settles it before noon. Then they start the run by hand, and the test is green again. The monthly report lists the case with its cause and the decision taken.
Test, alert, follow up, report
The work runs as a cycle: Maester checks the configuration, Defender reports attacks, and both reach a person who decides. How often the checks run we agree with you, for example weekly or monthly and also after every larger change to the tenant. It starts with a first run against your Zero Trust rules and the recommendations that fit your tenant; deviations you want on purpose are recorded as justified exceptions.
- What Maester checksMaester is an open source test framework for PowerShell, built on Pester and Microsoft Graph. Each test compares one setting in your tenant with the expected value, from Entra ID and conditional access through Exchange Online, Teams and SharePoint to Intune, Defender and Purview.Result: every setting that matters to you has a test with a date and a result.
- Test suites and your own testsBesides the Maester tests, the project ships well-known suites: EIDSCA for Entra ID, the SCuBA baselines of the US agency CISA, controls from the CIS Benchmarks and ORCA for Exchange Online. We choose with you what fits your tenant and write tests for your own rules, for example for emergency accounts, guest access or agents with their own identity.Result: the checks cover what applies in your organisation.
- Scheduled runThe tests run on a schedule in GitHub Actions or Azure DevOps, or in Azure Automation or as a container job in Azure, and can be started by hand at any time after a change. They sign in through an app registration with read permissions and workload identity federation, with no stored secret.Result: the run depends on no personal account and no laptop.
- Report and alertEvery run produces an HTML report with each test, its result and a description of what it checks, usually with a hint on how a deviation can be fixed; the same results are available as Markdown, JSON, CSV or Excel. When a test fails, a summary arrives by email or as a card in a Teams channel, with a link to the report.Result: a deviation reaches the person who can assess it.
- Follow-upEvery failed test becomes an entry with an owner, a decision and a deadline: fix it, record it as a deliberate exception or adapt the test to a new rule.Result: every red test gets a decision, and every exception carries a reason.
- Signals from MicrosoftSecure Score shows which recommendations are open and how the score develops; incidents in Defender XDR go by email, filtered by severity, to the people who respond.Result: configuration and attacks come together in one overview.
- Monthly report and rolesOne report a month with test results, the Secure Score trend, incidents and open decisions, plus a table of who responds to which alert and who stands in.Result: management and IT see on one page where the tenant stands.
In place before the first run
Your tenant's Zero Trust state as the starting point, with the rhythm you want for the runs, the rules that should apply and the exceptions you make on purpose. A repository in GitHub or Azure DevOps and an app registration whose permissions your IT approves. A mailbox or a Teams channel for the alerts. And for each kind of alert a person who decides, with a deputy.
What runs after setup
Scheduled runs at the agreed rhythm, active in your repository, with the first report and the exceptions you have confirmed. Alerts by email or Teams, checked with a test that fails on purpose. Notification rules for Defender incidents, a template for the monthly report and the roles table. Whether your IT handles the alerts itself or Support & Operations does is set out in that table.
Sound familiar? Tell us where you stand today.
ContactWhat's new.
As of: 9 October 2026
- As of 9 October 2026
The Maester test suites, version 2.3.
The tests overview on maester.dev lists the Maester tests for Microsoft 365 and Entra, Entra ID SCA (EIDSCA), the CISA SCuBA baselines, controls from the CIS Benchmarks, ORCA for Exchange Online and a suite for on-premises Active Directory. The project is on GitHub under the MIT licence; results can be exported as HTML, Markdown, JSON, CSV or Excel and sent by email, Teams or Slack.
Source: maester.dev: Tests overview - Since 23 September 2026
Integrated Security Operations Center in Microsoft Defender, in preview.
ISOC brings XDR, SIEM, threat intelligence, automation and AI into the Defender portal. Since 23 September 2026 the preview is open to eligible Microsoft 365 E5 and E7 customers without an active Microsoft Sentinel workspace.
Source: Microsoft Learn: What's new in Microsoft Defender XDR - Since April 2026
Built-in rules hide benign alerts.
The built-in alert tuning rules are generally available. They suppress alerts from common benign activity in Defender for Endpoint and Defender for Office 365; automated investigations and email notifications are not affected.
Source: Microsoft Learn: What's new in Microsoft Defender XDR - As of 7 March 2026
Secure Score updates continuously, Entra recommendations weekly.
The score updates in real time and syncs system data once a day. Microsoft re-evaluates Entra ID recommendations on every change and also weekly, Teams recommendations monthly. Microsoft states that the score is no guarantee against a breach.
Source: Microsoft Learn: Microsoft Secure Score
How Maester runs for you.
The tests are the same whether your IT starts them, a pipeline schedules them or a portal shows their history. Which route fits your tenant we decide with you before the first run.
- PowerShell
Locally in PowerShell
The module is installed on an admin device, signs in to Microsoft Graph and produces an HTML report. This fits a first assessment and a run started by hand after a change.
- Pipeline
Scheduled in your cloud
GitHub Actions, Azure DevOps, Azure Automation or a container job in Azure start the tests on a schedule, signed in with no stored secret. Report and alert go to the agreed mailbox or Teams channel.
- Portal
Maester Cloud
A portal built on Maester with history over time, drift between two runs, alerts on changes and a view across several tenants. Today there is a self-hosted edition in your own Azure subscription, sold as a subscription; a hosted edition has been announced.
maester.cloud
What sits behind the alerts.
The products whose settings are checked regularly, and a tool for a quick check in between.
Entra, Defender, Intune and Purview at a glance.
Entra IDIdentities, roles and conditional access, checked on every run.
Copilot Studio agentsAgents with their own identity in Entra ID that belong in the checks.
Business or EnterpriseWhich Microsoft 365 plans include which Defender features.
SharePoint CheckerCheck sharing and site settings once against the recommendations.
Security resourcesArticles and guides on security in Microsoft 365.
Checking agents continuouslyLogs, tests and a rehearsed emergency stop once agents work in production.
Checked regularly, in our projects.
Recurring configuration checks with history at BMC, and a shared security baseline at SAGW against which later changes can be measured.
Related to your task
Sign-in, conditional access and time-limited admin rights that Maester tests on every run.
Microsoft Defender rollout & operationsHandle and close the incidents Defender reports, together with your IT.
Compliance & Data ProtectionConfidentiality, data loss prevention and retention under the FADP.
Microsoft 365 governance with PurviewOwnership, lifecycle and sharing for sites and teams.
Support & OperationsHave alerts handled with agreed response times.
Before the first test runs.
Does Maester change anything in our tenant?
The tests read settings and compare them with what is expected; nothing is changed. The app registration gets read permissions in Microsoft Graph and a role for viewing the configuration in Exchange Online. Some SharePoint tests need wider rights; whether we enable them, you decide with us before the first run.
Who is behind Maester?
The Maester Team, as maester.dev names its publishers, maintains the project with the community on GitHub, as open source under the MIT licence. It is built on the PowerShell test framework Pester and on Microsoft Graph. The Maester Team keeps adding new tests. Every test is open in the source code, and we write tests for your own rules in the same form.
How often do the tests run?
As often as your tenant changes. For many organisations a weekly or monthly run fits, plus a run started by hand after every larger change, such as a new conditional access policy. We set the rhythm in the schedule and adjust it when your tenant changes faster or slower. Maester checks the configuration; attacks between two runs are reported by Defender XDR with its own notifications.
Does Maester need a licence of its own?
No, Maester is open source under the MIT licence. What the tests can check depends on your Microsoft 365 plans: tests for features your plan does not include, such as Entra ID P2, are shown in the report as skipped. Which plans include which Defender features is set out in the comparison of Business and Enterprise.
Why do the tests run in GitHub Actions or Azure DevOps?
Because schedule, log and history are already there: every run has a date, a result and a link that every alert carries along. Your own tests sit in the same repository and are reviewed like code. If you use neither, Maester runs in Azure Automation or as a container job in Azure.
How do we read Secure Score?
As a measure of how many recommended protections are in place, with a trend and a comparison with similar organisations. Microsoft states that the score is no guarantee against a breach. In the monthly report, what counts is the direction and every recommendation you mark as an accepted risk on purpose.
Can we read the reports ourselves?
Yes. The test report is an HTML page listing every test, its result and a description of what it checks; the monthly report sums this up for management. We go through the first reports with your IT until it can place deviations on its own.
Agree on a useful next step
Tell us about your current workflow, the systems involved and the result you need. In our first conversation, we identify what else we need to know to plan a useful next step.


