Microsoft 365 should have a tuned anti-phishing policy that covers your own domains and key people
This check looks at the anti-phishing policies in Microsoft Defender for Office 365 and whether one of them is set up with impersonation and mailbox-intelligence protection instead of only the built-in defaults. The safe setting is a custom, enabled policy applied to your accepted domains, with impersonation protection and spoof intelligence turned on.
Why it matters
Phishing mail that pretends to be your CEO, your finance team or one of your own domains is one of the most common ways an attacker gets a first foothold in a tenant. Microsoft ships a default anti-phishing policy, but it cannot know which of your people are worth impersonating or which partner domains you trust. A policy you create and tune yourself closes that gap: you name the users and domains to protect, choose how aggressive detection should be, and decide what happens to a message that trips it.
Microsoft describes the available protections, such as spoof intelligence, mailbox intelligence and impersonation protection, in its anti-phishing documentation. Reading it before you change anything helps you pick the thresholds that suit your mail flow.
Source: Anti-phishing policies in Microsoft 365 (Microsoft Learn).
How to check it yourself
Portal: open the Microsoft Defender portal, then Email & collaboration → Policies & rules → Threat policies → Anti-phishing. Look for a policy besides the default one, confirm it is turned on and see which users, groups or domains it is applied to.
PowerShell (read-only): connect to Exchange Online and list the settings. These cmdlets only read.
Connect-ExchangeOnline
Get-AntiPhishPolicy | Format-List Name, IsDefault, Enabled, PhishThresholdLevel,
EnableSpoofIntelligence, EnableMailboxIntelligence, EnableMailboxIntelligenceProtection,
EnableTargetedUserProtection, EnableOrganizationDomainsProtection
Get-AntiPhishRule | Format-Table Name, State, Priority, RecipientDomainIsA tenant that shows only the default policy, or a custom policy whose impersonation settings are all False, does not meet the safe setting. Run this from a test session first if you are unsure about your permissions; no object is modified.
Reference: Get-AntiPhishPolicy (Microsoft Learn).
How to fix it
Portal: in the Defender portal go to Threat policies → Anti-phishing → Create. Name the policy, apply it to your accepted domains, then in the impersonation section add the users who are most likely to be imitated (executives, finance, HR) and the partner domains you deal with most. Turn on mailbox intelligence and spoof intelligence, choose an action for detected impersonation (quarantine is a common starting point) and save.
PowerShell: the example below creates a policy and a rule. It changes your tenant when you run it, so try it on a non-production tenant or with -WhatIf first. ConfigCobra itself never makes this change; it only reports the finding and suggests the fix.
New-AntiPhishPolicy -Name "Org anti-phishing" -Enabled $true -PhishThresholdLevel 2 `
-EnableSpoofIntelligence $true -EnableMailboxIntelligence $true `
-EnableMailboxIntelligenceProtection $true -EnableOrganizationDomainsProtection $true `
-EnableTargetedUserProtection $true -TargetedUsersToProtect "Pat Example;pat@contoso.example"
New-AntiPhishRule -Name "Org anti-phishing" -AntiPhishPolicy "Org anti-phishing" `
-RecipientDomainIs "contoso.example" -Priority 0Reference: Configure anti-phishing policies (Microsoft Learn).
What changes when you fix it
Recipients will see some inbound messages go to quarantine or arrive with a safety tip, depending on the action you chose. The first days are the noisy ones: legitimate mail from look-alike partner domains or from a vendor that really does send on behalf of your executives can be caught.
Help desk and mail admins should expect a few release requests and should review the quarantine daily at first. Adding trusted senders through the policy's own settings, instead of broad allow lists, keeps the protection intact.
Nothing else changes: mailboxes, mail routing and outbound mail are untouched. Roll out in stages if you can: apply the policy to a pilot group, watch the quarantine, then widen the scope to all accepted domains. Choose a lower threshold first and raise it once the false-positive rate is acceptable.
Included in
| Benchmark | CIS ID | Profile |
|---|---|---|
| CIS Microsoft 365 Foundations v7.0.0 | 2.1.7 | E5 Level 2 |
More checks in Microsoft Defender
- Safe Links for Office Applications should be enabledCIS 2.1.1
- The common attachment types filter should be enabledCIS 2.1.2
- Notifications for internal users sending malware should be enabledCIS 2.1.3
- Safe Attachments policy should be enabledCIS 2.1.4
- Safe Attachments for SharePoint, OneDrive, and Microsoft Teams should be enabledCIS 2.1.5
Go further
- Microsoft Defender21 controls in this section
- CIS Microsoft 365 Foundations v7.0.0160 controls
- CIS Microsoft 365 Foundations checklistEvery control in one list, as CSV
- RemediationSee what a fix changes in your tenant before you apply it, then use the generated script
- Monitoring & DriftScheduled scans, drift detection and email alerts when a control regresses
Check this on your Microsoft 365
ConfigCobra runs this check with every other control of the CIS Microsoft 365 Foundations Benchmark, read-only.