Skip to content

Microsoft 365 checks

Trimtab runs these 57 read-only checks against each connected Microsoft 365 tenant. Checks that need Microsoft Entra ID P1 or Exchange Online access show as not evaluated when that isn't available. See Connect Microsoft 365.

ID Check Severity Standards
M365-001 MFA is enforced for all users High CMMC L2, HIPAA, CMMC L1, CIS IG1, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
M365-002 MFA is enforced for administrators High CMMC L2, HIPAA, CMMC L1, CIS IG1, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
M365-003 Legacy authentication is blocked High CMMC L2, HIPAA, CMMC L1, CIS IG1, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
M365-004 Users are registered for MFA Medium CMMC L2, HIPAA, CMMC L1, CIS IG1, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
M365-005 Global Administrator count is controlled High CMMC L2, HIPAA, CMMC L1, CIS IG1, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
M365-006 Users cannot consent to risky apps Medium CMMC L2, HIPAA, CMMC L1, CIS IG1, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
M365-007 Users cannot register applications Low CMMC L2, CMMC L1, CIS IG1, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
M365-008 Guest access and invitations are restricted Medium CMMC L2, HIPAA, CMMC L1, CIS IG1, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
M365-009 SharePoint and OneDrive do not allow anonymous links High CMMC L2, HIPAA, CMMC L1, CIS IG1, FTC, NIST CSF, NYDFS, Reg S-P, SOC 2, Insurance
M365-010 No stale enabled accounts Medium CMMC L2, HIPAA, CMMC L1, CIS IG1, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
M365-011 Microsoft Secure Score meets target Low CMMC L2, HIPAA, CIS IG1, FTC, NIST CSF, NYDFS, Reg S-P, SOC 2
M365-012 Idle sessions in Microsoft 365 web apps sign out Medium CMMC L2, HIPAA, FTC, NIST CSF, NYDFS, Reg S-P, SOC 2, Insurance
EXO-001 Mailbox auditing is on for the organization High CMMC L2, HIPAA, CIS IG1, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
EXO-002 Unified audit log is on High CMMC L2, HIPAA, CIS IG1, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
EXO-003 Automatic forwarding to external addresses is blocked High CMMC L2, HIPAA, CMMC L1, CIS IG1, FTC, NIST CSF, NYDFS, Reg S-P, SOC 2, Insurance
EXO-004 No mailboxes forward to external addresses Medium CMMC L2, HIPAA, CMMC L1, CIS IG1, FTC, NIST CSF, NYDFS, Reg S-P, SOC 2, Insurance
EXO-005 Modern authentication is on for Exchange Medium CMMC L2, HIPAA, CMMC L1, CIS IG1, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
EXO-006 SMTP AUTH is turned off Medium CMMC L2, HIPAA, CMMC L1, CIS IG1, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
EXO-007 DKIM signing is on for custom domains Medium CMMC L2, HIPAA, CMMC L1, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
EXO-008 DMARC is enforced for custom domains Medium CMMC L2, HIPAA, CMMC L1, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
EXO-009 Malware filter blocks risky attachment types Medium CMMC L2, HIPAA, CMMC L1, CIS IG1, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
EXO-010 SPF is published for custom domains Medium CMMC L2, HIPAA, CMMC L1, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
EXO-011 Anti-phishing protection is on Medium CMMC L2, HIPAA, CMMC L1, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
M365-013 Admin accounts are cloud-only Medium CMMC L2, CMMC L1, HIPAA, CIS IG1, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
M365-014 Risky sign-ins and risky users are challenged Medium CMMC L2, HIPAA, FTC, NIST CSF, Reg S-P, SOC 2, Insurance
M365-015 Sign-ins require a managed device Medium CMMC L2, CMMC L1, HIPAA, FTC, NIST CSF, NYDFS, Reg S-P, SOC 2, Insurance
M365-016 Text and voice-call sign-in codes are turned off Low CMMC L2, HIPAA, FTC, NIST CSF, PCI DSS, NYDFS, SOC 2, Insurance
M365-017 Admin consent requests are turned on Low CMMC L2, FTC, NIST CSF, SOC 2
M365-018 Users can't create new tenants Low CMMC L2, FTC, NIST CSF, SOC 2
M365-019 No stale guest accounts Low CMMC L2, HIPAA, CIS IG1, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
SPO-001 SharePoint blocks legacy authentication Medium CMMC L2, CMMC L1, HIPAA, CIS IG1, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
SPO-002 Guests can't reshare files Low CMMC L2, CMMC L1, HIPAA, CIS IG1, FTC, NIST CSF, Reg S-P, SOC 2
SPO-003 OneDrive syncs only on managed devices Low CMMC L2, HIPAA, FTC, NIST CSF, NYDFS, Reg S-P, SOC 2
EXO-012 Safe Links and Safe Attachments are on Medium CMMC L2, CMMC L1, HIPAA, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance, CIS IG1
EXO-013 Mail flow rules don't skip spam filtering for outside senders Medium CMMC L2, CMMC L1, HIPAA, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
EXO-014 Spam filters don't allow-list senders or IP addresses Medium CMMC L2, CMMC L1, HIPAA, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
EXO-015 Email from outside the organization is tagged Low HIPAA, FTC, NIST CSF, PCI DSS, SOC 2, Insurance
EXO-016 Shared mailboxes can't be signed in to Medium CMMC L2, CMMC L1, HIPAA, FTC, NIST CSF, PCI DSS, Reg S-P, SOC 2, Insurance
EXO-017 Users can't install Outlook add-ins Low CMMC L2, FTC, NIST CSF, SOC 2, Insurance
M365-020 Admins use phishing-resistant MFA Medium CMMC L2, CMMC L1, HIPAA, CIS IG1, FTC, NIST CSF, PCI DSS, NYDFS, SOC 2, Insurance
M365-021 Global Administrators use just-in-time access Medium CMMC L2, HIPAA, CIS IG1, FTC, NIST CSF, PCI DSS, NYDFS, SOC 2, Insurance
M365-022 Emergency access accounts are set up Low FTC, NIST CSF, SOC 2, Insurance
M365-023 Device code sign-in is blocked Medium CMMC L2, HIPAA, FTC, NIST CSF, Reg S-P, SOC 2, Insurance
M365-024 Admin sessions re-authenticate and don't stay signed in Low CMMC L2, HIPAA, CIS IG1, FTC, NIST CSF, PCI DSS, SOC 2, Insurance
M365-025 Company-specific banned passwords are on Low CMMC L2, HIPAA, CIS IG1, FTC, NIST CSF, PCI DSS, NYDFS, SOC 2, Insurance
M365-026 Trimtab's own access is read-only High CMMC L2, CMMC L1, HIPAA, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2
SPO-004 SharePoint sharing is limited to approved domains Low CMMC L2, CMMC L1, HIPAA, FTC, NIST CSF, Reg S-P, SOC 2
EXO-018 Calendar details aren't shared outside the company Low CMMC L2, HIPAA, FTC, NIST CSF, Reg S-P, SOC 2
EXO-019 Admins are notified about outbound spam Low FTC, NIST CSF, Reg S-P, SOC 2, Insurance
EXO-020 Zero-hour auto purge removes threats after delivery Low CMMC L2, CMMC L1, HIPAA, FTC, NIST CSF, PCI DSS, NYDFS, Reg S-P, SOC 2, Insurance
INT-001 Every device platform in use has a compliance policy Medium CMMC L2, CIS IG1, FTC, NIST CSF, NYDFS, SOC 2, Insurance
INT-002 Devices without a compliance policy count as not compliant Medium CMMC L2, CIS IG1, FTC, NIST CSF, SOC 2
INT-003 Managed devices are compliant Medium CMMC L2, HIPAA, CIS IG1, FTC, NIST CSF, NYDFS, SOC 2, Insurance
INT-004 Intune-managed devices are encrypted High CMMC L2, HIPAA, CIS IG1, FTC, NIST CSF, NYDFS, Reg S-P, SOC 2, Insurance
INT-005 Compliance policies require a passcode and encryption Medium CMMC L2, HIPAA, CIS IG1, FTC, NIST CSF, NYDFS, SOC 2, Insurance
INT-006 App protection policies protect company data on phones Medium CMMC L2, HIPAA, FTC, NIST CSF, Reg S-P, SOC 2
INT-007 Managed devices check in regularly Low CMMC L2, CIS IG1, FTC, NIST CSF, NYDFS, PCI DSS, SOC 2

Identity

MFA is enforced for all users (M365-001)

High Identity

Fix: Turn on security defaults, or create a Conditional Access policy that requires MFA for all users and all cloud apps.

Take the helm: step-by-step
  1. Without Entra ID P1: Entra admin center → Identity → Overview → Properties → Manage security defaults → Enabled.
  2. With P1: Protection → Conditional Access → New policy. Users = All users (exclude two emergency access accounts), Target resources = All cloud apps, Grant = Require multifactor authentication.
  3. Run it in Report-only for a few days, then switch it On.
  4. Ask users to register an authenticator app beforehand so nobody gets locked out.

Maps to:

  • IA.L2-3.5.3 Use multifactor authentication. Requiring MFA for everyone means a stolen password alone can't get into Microsoft 365.
  • 164.312(d) Person or entity authentication. Requiring MFA for everyone means a stolen password alone can't get into Microsoft 365.
  • IA.L1-B.1.VI Authenticate users and devices. Requiring MFA for everyone means a stolen password alone can't get into Microsoft 365.
  • 6.3 Require MFA for Externally-Exposed Applications. Requiring MFA for everyone means a stolen password alone can't get into Microsoft 365.
  • 314.4(c)(5) Multifactor authentication. Requiring MFA for everyone means a stolen password alone can't get into Microsoft 365.
  • PR.AA-03 Users authenticated. Requiring MFA for everyone means a stolen password alone can't get into Microsoft 365.
  • 8.4.2 MFA for access into the CDE. Requiring MFA for everyone means a stolen password alone can't get into Microsoft 365.
  • 8.4.3 MFA for remote access. Requiring MFA for everyone means a stolen password alone can't get into Microsoft 365.
  • 500.12(a) Multi-factor authentication. Requiring MFA for everyone means a stolen password alone can't get into Microsoft 365.
  • 248.30(a)(2)(iii) Protection against unauthorized access. Requiring MFA for everyone means a stolen password alone can't get into Microsoft 365.
  • CC6.1 Logical access security. Requiring MFA for everyone means a stolen password alone can't get into Microsoft 365.
  • CI-01 MFA for email and cloud apps. Requiring MFA for everyone means a stolen password alone can't get into Microsoft 365.

MFA is enforced for administrators (M365-002)

High Identity

Fix: Require MFA for all directory roles with a Conditional Access policy, or turn on security defaults.

Take the helm: step-by-step
  1. Security defaults already require MFA for admins.
  2. With P1: Conditional Access → New policy. Users → Directory roles = Global Administrator and the other admin roles, Target resources = All cloud apps, Grant = Require multifactor authentication (or phishing-resistant MFA). Turn it On.

Maps to:

  • IA.L2-3.5.3 Use multifactor authentication. Admins hold the keys to the tenant. MFA on admin accounts stops a stolen password from becoming a full takeover.
  • AC.L2-3.1.5 Employ the principle of least privilege. Extra protection on privileged accounts is part of keeping admin access tight and deliberate.
  • 164.312(d) Person or entity authentication. Admins hold the keys to the tenant. MFA on admin accounts stops a stolen password from becoming a full takeover.
  • IA.L1-B.1.VI Authenticate users and devices. Admins hold the keys to the tenant. MFA on admin accounts stops a stolen password from becoming a full takeover.
  • 6.5 Require MFA for Administrative Access. Admins hold the keys to the tenant. MFA on admin accounts stops a stolen password from becoming a full takeover.
  • 314.4(c)(5) Multifactor authentication. Admins hold the keys to the tenant. MFA on admin accounts stops a stolen password from becoming a full takeover.
  • PR.AA-03 Users authenticated. Admins hold the keys to the tenant. MFA on admin accounts stops a stolen password from becoming a full takeover.
  • PR.AA-05 Least privilege. Admins hold the keys to the tenant. MFA on admin accounts stops a stolen password from becoming a full takeover.
  • 8.4.2 MFA for access into the CDE. Admins hold the keys to the tenant. MFA on admin accounts stops a stolen password from becoming a full takeover.
  • 500.12(a) Multi-factor authentication. Admins hold the keys to the tenant. MFA on admin accounts stops a stolen password from becoming a full takeover.
  • 500.7(a)(2) Limit privileged accounts. Admins hold the keys to the tenant. MFA on admin accounts stops a stolen password from becoming a full takeover.
  • 248.30(a)(2)(iii) Protection against unauthorized access. Admins hold the keys to the tenant. MFA on admin accounts stops a stolen password from becoming a full takeover.
  • CC6.1 Logical access security. Admins hold the keys to the tenant. MFA on admin accounts stops a stolen password from becoming a full takeover.
  • CC6.3 Least privilege. Admins hold the keys to the tenant. MFA on admin accounts stops a stolen password from becoming a full takeover.
  • CI-02 MFA and limits on admin accounts. Admins hold the keys to the tenant. MFA on admin accounts stops a stolen password from becoming a full takeover.

Legacy authentication is blocked (M365-003)

High Identity

Fix: Block legacy authentication (Exchange ActiveSync and other clients) with Conditional Access, or turn on security defaults.

Take the helm: step-by-step
  1. Security defaults already block legacy authentication.
  2. With P1, first check Sign-in logs (filter Client app = the legacy clients) for anything still using it.
  3. Conditional Access → New policy. Users = All users, Target resources = All cloud apps, Conditions → Client apps = Exchange ActiveSync clients and Other clients, Grant = Block access.
  4. Run it in Report-only first, then switch it On.

Maps to:

  • IA.L2-3.5.3 Use multifactor authentication. Old sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.
  • CM.L2-3.4.7 Restrict nonessential programs, functions, ports, protocols, and services. Legacy authentication is an outdated protocol that should be turned off.
  • 164.312(d) Person or entity authentication. Old sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.
  • IA.L1-B.1.VI Authenticate users and devices. Old sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.
  • 6.3 Require MFA for Externally-Exposed Applications. Old sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.
  • 314.4(c)(5) Multifactor authentication. Old sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.
  • PR.AA-03 Users authenticated. Old sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.
  • 8.4.2 MFA for access into the CDE. Old sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.
  • 2.2.5 Insecure services justified and secured. Old sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.
  • 500.12(a) Multi-factor authentication. Old sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.
  • 248.30(a)(2)(iii) Protection against unauthorized access. Old sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.
  • CC6.6 Protection from outside threats. Old sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.
  • CI-01 MFA for email and cloud apps. Old sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.

Users are registered for MFA (M365-004)

Medium Identity

Fix: Have remaining users register an authentication method, for example with a registration campaign.

Default thresholds: Target percentage = 95

Take the helm: step-by-step
  1. See who isn't registered: Entra admin center → Protection → Authentication methods → User registration details.
  2. Turn on a registration campaign: Authentication methods → Registration campaign → Enabled for all users.
  3. Follow up individually with anyone still not registered.

Maps to:

  • IA.L2-3.5.3 Use multifactor authentication. MFA only protects people who have set it up. This checks that nearly everyone has.
  • 164.312(d) Person or entity authentication. MFA only protects people who have set it up. This checks that nearly everyone has.
  • IA.L1-B.1.VI Authenticate users and devices. MFA only protects people who have set it up. This checks that nearly everyone has.
  • 6.3 Require MFA for Externally-Exposed Applications. MFA only protects people who have set it up. This checks that nearly everyone has.
  • 314.4(c)(5) Multifactor authentication. MFA only protects people who have set it up. This checks that nearly everyone has.
  • PR.AA-03 Users authenticated. MFA only protects people who have set it up. This checks that nearly everyone has.
  • 8.4.2 MFA for access into the CDE. MFA only protects people who have set it up. This checks that nearly everyone has.
  • 500.12(a) Multi-factor authentication. MFA only protects people who have set it up. This checks that nearly everyone has.
  • 248.30(a)(2)(iii) Protection against unauthorized access. MFA only protects people who have set it up. This checks that nearly everyone has.
  • CC6.1 Logical access security. MFA only protects people who have set it up. This checks that nearly everyone has.
  • CI-01 MFA for email and cloud apps. MFA only protects people who have set it up. This checks that nearly everyone has.

No stale enabled accounts (M365-010)

Medium Identity

Fix: Disable or remove accounts that have not signed in within the inactivity period.

Default thresholds: Inactivity period (days) = 90

Take the helm: step-by-step
  1. Review the inactive accounts listed in the evidence.
  2. Confirm with the manager or HR, then block sign-in (Users → user → Edit properties → Account enabled: off) or delete the account.
  3. Set up access reviews (Entra ID P2) or a monthly review so inactive accounts get caught.

Maps to:

  • IA.L2-3.5.6 Disable identifiers after a period of inactivity. Unused accounts are easy targets because nobody notices when they're broken into. Disabling them removes that risk.
  • 164.312(a)(1) Access control. Unused accounts are easy targets because nobody notices when they're broken into. Disabling them removes that risk.
  • AC.L1-B.1.I Limit access to authorized users. Unused accounts are easy targets because nobody notices when they're broken into. Disabling them removes that risk.
  • 5.3 Disable Dormant Accounts. Unused accounts are easy targets because nobody notices when they're broken into. Disabling them removes that risk.
  • 314.4(c)(1) Access controls. Unused accounts are easy targets because nobody notices when they're broken into. Disabling them removes that risk.
  • PR.AA-01 Identities and credentials managed. Unused accounts are easy targets because nobody notices when they're broken into. Disabling them removes that risk.
  • 8.2.6 Inactive accounts removed. Unused accounts are easy targets because nobody notices when they're broken into. Disabling them removes that risk.
  • 500.7(a)(4) Remove unneeded accounts. Unused accounts are easy targets because nobody notices when they're broken into. Disabling them removes that risk.
  • 248.30(a)(2)(iii) Protection against unauthorized access. Unused accounts are easy targets because nobody notices when they're broken into. Disabling them removes that risk.
  • CC6.2 Granting and removing access. Unused accounts are easy targets because nobody notices when they're broken into. Disabling them removes that risk.
  • CI-12 Account and password hygiene. Unused accounts are easy targets because nobody notices when they're broken into. Disabling them removes that risk.

Idle sessions in Microsoft 365 web apps sign out (M365-012)

Medium Identity

Fix: Turn on idle session timeout for Microsoft 365 web apps (Microsoft 365 admin center, Org settings, Security & privacy).

Default thresholds: Maximum idle time (hours) = 3

Take the helm: step-by-step
  1. Microsoft 365 admin center → Settings → Org settings → Security & privacy → Idle session timeout.
  2. Turn on 'Turn on to set the period of inactivity for users to be signed off of Microsoft 365 web apps' and choose 3 hours or less. 1 hour suits most offices.
  3. Users get a warning before they're signed out. It applies to Outlook on the web, SharePoint, OneDrive, and the other Microsoft 365 web apps.
  4. It doesn't sign out desktop apps or devices. Screen lock (SES-001) covers computers.

Maps to:

  • AC.L2-3.1.11 Terminate user sessions after a defined condition. Signing out idle web sessions means a browser left open on a shared or lost computer doesn't stay signed in to email and files.
  • 164.312(a)(2)(iii) Automatic logoff. Signing out idle web sessions means a browser left open on a shared or lost computer doesn't stay signed in to email and files.
  • 314.4(c)(1) Access controls. Signing out idle web sessions means a browser left open on a shared or lost computer doesn't stay signed in to email and files.
  • PR.AA-03 Users authenticated. Signing out idle web sessions means a browser left open on a shared or lost computer doesn't stay signed in to email and files.
  • 500.3(g) Systems and network security. Signing out idle web sessions means a browser left open on a shared or lost computer doesn't stay signed in to email and files.
  • 248.30(a)(2)(iii) Protection against unauthorized access. Signing out idle web sessions means a browser left open on a shared or lost computer doesn't stay signed in to email and files.
  • CC6.1 Logical access security. Signing out idle web sessions means a browser left open on a shared or lost computer doesn't stay signed in to email and files.
  • CI-12 Account and password hygiene. Carriers ask about session timeouts. Signing out idle web sessions keeps abandoned sign-ins from being reused.

Risky sign-ins and risky users are challenged (M365-014)

Medium Identity

Fix: With Entra ID P2, add Conditional Access policies that require MFA for medium and high sign-in risk and a secure password change for high user risk.

Take the helm: step-by-step
  1. Needs Microsoft Entra ID P2 (included in Microsoft 365 E5).
  2. Entra admin center → Protection → Conditional Access → New policy: Users = All users (exclude emergency access accounts), Target resources = All cloud apps, Conditions → Sign-in risk = High and Medium, Grant = Require multifactor authentication.
  3. Another policy: Conditions → User risk = High, Grant = Require password change (with MFA).
  4. Run both in Report-only for a few days, then switch them On.

Maps to:

  • AC.L2-3.1.1 Limit system access to authorized users. Microsoft flags sign-ins and accounts that look compromised. Challenging them stops an attacker using a stolen password even when it's correct.
  • 164.308(a)(5)(ii)(C) Log-in monitoring. Risk-based policies watch every sign-in for signs of an attack and respond automatically.
  • 314.4(c)(8) Monitor and log user activity. Microsoft flags sign-ins and accounts that look compromised. Challenging them stops an attacker using a stolen password even when it's correct.
  • DE.CM-03 Personnel activity monitored. Microsoft watches sign-in activity for unusual behavior, and these policies act on what it finds.
  • 248.30(a)(2)(ii) Protection against anticipated threats. Microsoft flags sign-ins and accounts that look compromised. Challenging them stops an attacker using a stolen password even when it's correct.
  • CC6.6 Protection from outside threats. Microsoft flags sign-ins and accounts that look compromised. Challenging them stops an attacker using a stolen password even when it's correct.
  • CC7.2 Monitoring for attacks. Microsoft flags sign-ins and accounts that look compromised. Challenging them stops an attacker using a stolen password even when it's correct.
  • CI-01 MFA for email and cloud apps. Microsoft flags sign-ins and accounts that look compromised. Challenging them stops an attacker using a stolen password even when it's correct.

Sign-ins require a managed device (M365-015)

Medium Identity

Fix: Enroll devices in Intune (or hybrid-join them), then add a Conditional Access policy that requires a compliant or hybrid-joined device for all users and all cloud apps.

Take the helm: step-by-step
  1. Enroll Windows, Mac, and mobile devices in Intune, or hybrid-join Windows computers to Entra ID, and create compliance policies.
  2. Entra admin center → Protection → Conditional Access → New policy: Users = All users (exclude emergency access accounts), Target resources = All cloud apps, Grant = Require device to be marked as compliant, or Require Microsoft Entra hybrid joined device.
  3. Run it in Report-only first and check the sign-in logs for devices that would be blocked, then switch it On.

Maps to:

  • AC.L2-3.1.1 Limit system access to authorized users. Requiring a managed device means a stolen password can't be used from an attacker's computer, only from devices you control.
  • AC.L2-3.1.12 Monitor and control remote access sessions. Requiring a managed device means a stolen password can't be used from an attacker's computer, only from devices you control.
  • AC.L1-B.1.I Limit access to authorized users. Requiring a managed device means a stolen password can't be used from an attacker's computer, only from devices you control.
  • IA.L1-B.1.V Identify users and devices. Each device has to be known to your management system before it can reach company data.
  • 164.312(a)(1) Access control. Requiring a managed device means a stolen password can't be used from an attacker's computer, only from devices you control.
  • 314.4(c)(2) Know your systems and devices. Only devices you know about and manage can reach customer information.
  • PR.AA-03 Users authenticated. Requiring a managed device means a stolen password can't be used from an attacker's computer, only from devices you control.
  • 500.3(c) Asset inventory, device management, and end of life. Access is limited to devices that are enrolled in device management.
  • 248.30(a)(2)(iii) Protection against unauthorized access. Requiring a managed device means a stolen password can't be used from an attacker's computer, only from devices you control.
  • CC6.1 Logical access security. Requiring a managed device means a stolen password can't be used from an attacker's computer, only from devices you control.
  • CC6.6 Protection from outside threats. Requiring a managed device means a stolen password can't be used from an attacker's computer, only from devices you control.
  • CI-08 Secured remote access. Requiring a managed device means a stolen password can't be used from an attacker's computer, only from devices you control.

Text and voice-call sign-in codes are turned off (M365-016)

Low Identity

Fix: In Entra authentication methods, turn off SMS and Voice call once users have the Microsoft Authenticator app or a security key.

Take the helm: step-by-step
  1. Make sure users have registered the Microsoft Authenticator app or a security key (Protection → Authentication methods → User registration details).
  2. Entra admin center → Protection → Authentication methods → Policies → SMS → Disable. Do the same for Voice call.
  3. If a few people truly can't use an app, enable SMS for a small group only, not all users.

Maps to:

  • IA.L2-3.5.3 Use multifactor authentication. Text and voice codes can be intercepted or redirected by SIM-swap scams. The Authenticator app and security keys are much harder to steal.
  • 164.312(d) Person or entity authentication. Text and voice codes can be intercepted or redirected by SIM-swap scams. The Authenticator app and security keys are much harder to steal.
  • 314.4(c)(5) Multifactor authentication. Text and voice codes can be intercepted or redirected by SIM-swap scams. The Authenticator app and security keys are much harder to steal.
  • PR.AA-03 Users authenticated. Text and voice codes can be intercepted or redirected by SIM-swap scams. The Authenticator app and security keys are much harder to steal.
  • 8.3.2 Authentication factors protected. Text and voice codes can be intercepted or redirected by SIM-swap scams. The Authenticator app and security keys are much harder to steal.
  • 500.12(a) Multi-factor authentication. Text and voice codes can be intercepted or redirected by SIM-swap scams. The Authenticator app and security keys are much harder to steal.
  • CC6.1 Logical access security. Text and voice codes can be intercepted or redirected by SIM-swap scams. The Authenticator app and security keys are much harder to steal.
  • CI-01 MFA for email and cloud apps. Text and voice codes can be intercepted or redirected by SIM-swap scams. The Authenticator app and security keys are much harder to steal.

Device code sign-in is blocked (M365-023)

Medium Identity

Fix: Add a Conditional Access policy that blocks the device code flow (Conditions, Authentication flows) for all users, with exceptions only for devices that need it.

Take the helm: step-by-step
  1. Check the sign-in logs for device code sign-ins (filter Authentication protocol = Device code) to find anything that legitimately uses it, such as meeting room devices.
  2. Entra admin center → Protection → Conditional Access → New policy: Users = All users (exclude the accounts that need it), Target resources = All cloud apps, Conditions → Authentication flows → Device code flow, Grant = Block access.
  3. Run it in Report-only first, then switch it On.

Maps to:

  • AC.L2-3.1.1 Limit system access to authorized users. Device code phishing tricks people into approving an attacker's sign-in on a real Microsoft page. Blocking the device code flow stops it.
  • 164.312(d) Person or entity authentication. Device code phishing tricks people into approving an attacker's sign-in on a real Microsoft page. Blocking the device code flow stops it.
  • 314.4(c)(5) Multifactor authentication. Device code phishing tricks people into approving an attacker's sign-in on a real Microsoft page. Blocking the device code flow stops it.
  • PR.AA-03 Users authenticated. Device code phishing tricks people into approving an attacker's sign-in on a real Microsoft page. Blocking the device code flow stops it.
  • 248.30(a)(2)(ii) Protection against anticipated threats. Device code phishing tricks people into approving an attacker's sign-in on a real Microsoft page. Blocking the device code flow stops it.
  • CC6.6 Protection from outside threats. Device code phishing tricks people into approving an attacker's sign-in on a real Microsoft page. Blocking the device code flow stops it.
  • CI-01 MFA for email and cloud apps. Device code phishing tricks people into approving an attacker's sign-in on a real Microsoft page. Blocking the device code flow stops it.

Company-specific banned passwords are on (M365-025)

Low Identity

Fix: In Entra password protection, turn on the custom banned password list and add your company, product, and location names.

Take the helm: step-by-step
  1. Needs Microsoft Entra ID P1.
  2. Entra admin center → Protection → Authentication methods → Password protection.
  3. Set Enforce custom list = Yes and add your company name, product names, office locations, and local sports teams, then Save.

Maps to:

  • IA.L2-3.5.7 Enforce minimum password complexity. A custom banned list stops passwords built from your company name or location, which attackers try first.
  • 164.308(a)(5)(ii)(D) Password management. A custom banned list stops passwords built from your company name or location, which attackers try first.
  • 5.2 Use Unique Passwords. A custom banned list stops passwords built from your company name or location, which attackers try first.
  • 314.4(c)(1) Access controls. A custom banned list stops passwords built from your company name or location, which attackers try first.
  • PR.AA-01 Identities and credentials managed. A custom banned list stops passwords built from your company name or location, which attackers try first.
  • 8.3.6 Password length and complexity. A custom banned list stops passwords built from your company name or location, which attackers try first.
  • 500.7(b) Password policy. A custom banned list stops passwords built from your company name or location, which attackers try first.
  • CC6.1 Logical access security. A custom banned list stops passwords built from your company name or location, which attackers try first.
  • CI-12 Account and password hygiene. A custom banned list stops passwords built from your company name or location, which attackers try first.

Privileged access

Global Administrator count is controlled (M365-005)

High Privileged access

Fix: Keep between two and four Global Administrators. Use less privileged roles for day-to-day administration.

Default thresholds: Minimum = 2, Maximum = 4

Take the helm: step-by-step
  1. Review assignments: Entra admin center → Roles & admins → Global Administrator.
  2. Move people who only need part of that access to narrower roles, such as User Administrator or Exchange Administrator.
  3. If there's only one Global Admin, add a second: a cloud-only emergency access account with a long password stored securely.

Maps to:

  • AC.L2-3.1.5 Employ the principle of least privilege. Few Global Administrators means fewer powerful accounts to steal, but at least two so you're never locked out.
  • 164.312(a)(1) Access control. Few Global Administrators means fewer powerful accounts to steal, but at least two so you're never locked out.
  • AC.L1-B.1.II Limit access to permitted functions. Few Global Administrators means fewer powerful accounts to steal, but at least two so you're never locked out.
  • 5.4 Restrict Administrator Privileges to Dedicated Administrator Accounts. Few Global Administrators means fewer powerful accounts to steal, but at least two so you're never locked out.
  • 314.4(c)(1) Access controls. Few Global Administrators means fewer powerful accounts to steal, but at least two so you're never locked out.
  • PR.AA-05 Least privilege. Few Global Administrators means fewer powerful accounts to steal, but at least two so you're never locked out.
  • 7.2.2 Least privilege. Few Global Administrators means fewer powerful accounts to steal, but at least two so you're never locked out.
  • 500.7(a)(2) Limit privileged accounts. Few Global Administrators means fewer powerful accounts to steal, but at least two so you're never locked out.
  • 248.30(a)(2)(iii) Protection against unauthorized access. Few Global Administrators means fewer powerful accounts to steal, but at least two so you're never locked out.
  • CC6.3 Least privilege. Few Global Administrators means fewer powerful accounts to steal, but at least two so you're never locked out.
  • CI-02 MFA and limits on admin accounts. Few Global Administrators means fewer powerful accounts to steal, but at least two so you're never locked out.

Admin accounts are cloud-only (M365-013)

Medium Privileged access

Fix: Give each admin a separate cloud-only account (name@tenant.onmicrosoft.com) for admin roles, and remove admin roles from accounts synced from on-premises Active Directory.

Take the helm: step-by-step
  1. List admins whose accounts sync from on-premises: the evidence names them.
  2. Create a separate cloud-only account for each person's admin work (for example admin.name@yourtenant.onmicrosoft.com) and register MFA on it.
  3. Assign the admin roles to the cloud-only account, then remove them from the synced account.

Maps to:

  • AC.L2-3.1.5 Employ the principle of least privilege. An admin account that syncs from on-premises Active Directory can be taken over from the office network. Cloud-only admin accounts keep an on-premises breach from reaching Microsoft 365.
  • AC.L2-3.1.6 Use non-privileged accounts for nonsecurity functions. A separate admin account means everyday work happens from an account without admin rights.
  • AC.L1-B.1.II Limit access to permitted functions. An admin account that syncs from on-premises Active Directory can be taken over from the office network. Cloud-only admin accounts keep an on-premises breach from reaching Microsoft 365.
  • 164.312(a)(1) Access control. An admin account that syncs from on-premises Active Directory can be taken over from the office network. Cloud-only admin accounts keep an on-premises breach from reaching Microsoft 365.
  • 5.4 Restrict Administrator Privileges to Dedicated Administrator Accounts. Dedicated cloud-only admin accounts are kept apart from the accounts people use every day.
  • 314.4(c)(1) Access controls. An admin account that syncs from on-premises Active Directory can be taken over from the office network. Cloud-only admin accounts keep an on-premises breach from reaching Microsoft 365.
  • PR.AA-05 Least privilege. An admin account that syncs from on-premises Active Directory can be taken over from the office network. Cloud-only admin accounts keep an on-premises breach from reaching Microsoft 365.
  • 7.2.2 Least privilege. An admin account that syncs from on-premises Active Directory can be taken over from the office network. Cloud-only admin accounts keep an on-premises breach from reaching Microsoft 365.
  • 500.7(a)(2) Limit privileged accounts. An admin account that syncs from on-premises Active Directory can be taken over from the office network. Cloud-only admin accounts keep an on-premises breach from reaching Microsoft 365.
  • 248.30(a)(2)(iii) Protection against unauthorized access. An admin account that syncs from on-premises Active Directory can be taken over from the office network. Cloud-only admin accounts keep an on-premises breach from reaching Microsoft 365.
  • CC6.3 Least privilege. An admin account that syncs from on-premises Active Directory can be taken over from the office network. Cloud-only admin accounts keep an on-premises breach from reaching Microsoft 365.
  • CI-02 MFA and limits on admin accounts. An admin account that syncs from on-premises Active Directory can be taken over from the office network. Cloud-only admin accounts keep an on-premises breach from reaching Microsoft 365.

Admins use phishing-resistant MFA (M365-020)

Medium Privileged access

Fix: Add a Conditional Access policy that requires the built-in Phishing-resistant MFA authentication strength (security keys, passkeys, Windows Hello, or certificates) for admin roles.

Take the helm: step-by-step
  1. Register a phishing-resistant method for each admin first: a FIDO2 security key, a passkey in Microsoft Authenticator, or Windows Hello for Business.
  2. Entra admin center → Protection → Conditional Access → New policy: Users → Directory roles = Global Administrator and the other admin roles (exclude emergency access accounts), Target resources = All cloud apps.
  3. Grant → Require authentication strength → Phishing-resistant MFA. Run it in Report-only, then switch it On.

Maps to:

  • IA.L2-3.5.3 Use multifactor authentication. Phishing-resistant MFA, such as security keys and passkeys, can't be relayed by fake sign-in pages, which defeat text codes and app prompts. Admin accounts are the top target.
  • IA.L1-B.1.VI Authenticate users and devices. Phishing-resistant MFA, such as security keys and passkeys, can't be relayed by fake sign-in pages, which defeat text codes and app prompts. Admin accounts are the top target.
  • 164.312(d) Person or entity authentication. Phishing-resistant MFA, such as security keys and passkeys, can't be relayed by fake sign-in pages, which defeat text codes and app prompts. Admin accounts are the top target.
  • 6.5 Require MFA for Administrative Access. Admin access gets the strongest form of MFA, one a phishing site can't capture.
  • 314.4(c)(5) Multifactor authentication. Phishing-resistant MFA, such as security keys and passkeys, can't be relayed by fake sign-in pages, which defeat text codes and app prompts. Admin accounts are the top target.
  • PR.AA-03 Users authenticated. Phishing-resistant MFA, such as security keys and passkeys, can't be relayed by fake sign-in pages, which defeat text codes and app prompts. Admin accounts are the top target.
  • 8.4.2 MFA for access into the CDE. Phishing-resistant MFA, such as security keys and passkeys, can't be relayed by fake sign-in pages, which defeat text codes and app prompts. Admin accounts are the top target.
  • 500.12(a) Multi-factor authentication. Phishing-resistant MFA, such as security keys and passkeys, can't be relayed by fake sign-in pages, which defeat text codes and app prompts. Admin accounts are the top target.
  • CC6.1 Logical access security. Phishing-resistant MFA, such as security keys and passkeys, can't be relayed by fake sign-in pages, which defeat text codes and app prompts. Admin accounts are the top target.
  • CI-02 MFA and limits on admin accounts. Phishing-resistant MFA, such as security keys and passkeys, can't be relayed by fake sign-in pages, which defeat text codes and app prompts. Admin accounts are the top target.

Global Administrators use just-in-time access (M365-021)

Medium Privileged access

Fix: With Entra ID P2, make Global Administrator assignments eligible in Privileged Identity Management, so people activate the role only when they need it. Keep permanent assignments for emergency access accounts only.

Default thresholds: maxPermanent = 2

Take the helm: step-by-step
  1. Needs Microsoft Entra ID P2 (included in Microsoft 365 E5, or as an add-on).
  2. Entra admin center → Identity governance → Privileged Identity Management → Microsoft Entra roles → Roles → Global Administrator.
  3. Add each admin as Eligible, then remove their Active assignment. Keep Active (permanent) assignments only for one or two emergency access accounts.
  4. In the role settings, require MFA and a justification on activation, and set a maximum activation time such as 4 hours.

Maps to:

  • AC.L2-3.1.5 Employ the principle of least privilege. With just-in-time access, admins hold Global Administrator only for the hours they need it, so a stolen admin account is usually powerless.
  • AC.L2-3.1.7 Prevent non-privileged users from executing privileged functions. Privileged functions need an explicit activation, which is logged, instead of being always on.
  • 164.312(a)(1) Access control. With just-in-time access, admins hold Global Administrator only for the hours they need it, so a stolen admin account is usually powerless.
  • 5.4 Restrict Administrator Privileges to Dedicated Administrator Accounts. With just-in-time access, admins hold Global Administrator only for the hours they need it, so a stolen admin account is usually powerless.
  • 314.4(c)(1) Access controls. With just-in-time access, admins hold Global Administrator only for the hours they need it, so a stolen admin account is usually powerless.
  • PR.AA-05 Least privilege. With just-in-time access, admins hold Global Administrator only for the hours they need it, so a stolen admin account is usually powerless.
  • 7.2.2 Least privilege. With just-in-time access, admins hold Global Administrator only for the hours they need it, so a stolen admin account is usually powerless.
  • 500.7(a)(2) Limit privileged accounts. With just-in-time access, admins hold Global Administrator only for the hours they need it, so a stolen admin account is usually powerless.
  • CC6.3 Least privilege. With just-in-time access, admins hold Global Administrator only for the hours they need it, so a stolen admin account is usually powerless.
  • CI-06 Limited admin rights. With just-in-time access, admins hold Global Administrator only for the hours they need it, so a stolen admin account is usually powerless.

Emergency access accounts are set up (M365-022)

Low Privileged access

Fix: Create one or two cloud-only emergency access (break-glass) Global Administrator accounts with long passwords or security keys, store them securely, and exclude them from Conditional Access policies.

Take the helm: step-by-step
  1. Create one or two cloud-only accounts (for example breakglass1@yourtenant.onmicrosoft.com) with the Global Administrator role.
  2. Give each a long random password or a FIDO2 security key, and store the credentials securely offline.
  3. Exclude these accounts from Conditional Access policies, and set up an alert for any sign-in to them.

Maps to:

  • 314.4(c)(1) Access controls. Emergency access accounts let you back in if a Conditional Access mistake or an MFA outage locks out every admin.
  • PR.AA-01 Identities and credentials managed. Emergency access accounts let you back in if a Conditional Access mistake or an MFA outage locks out every admin.
  • CC6.1 Logical access security. Emergency access accounts let you back in if a Conditional Access mistake or an MFA outage locks out every admin.
  • CI-02 MFA and limits on admin accounts. Emergency access accounts let you back in if a Conditional Access mistake or an MFA outage locks out every admin.

Admin sessions re-authenticate and don't stay signed in (M365-024)

Low Privileged access

Fix: Add a Conditional Access policy for admin roles with sign-in frequency of 4 hours or less and persistent browser session set to Never persistent.

Default thresholds: Maximum idle time (hours) = 4

Take the helm: step-by-step
  1. Entra admin center → Protection → Conditional Access → New policy: Users → Directory roles = Global Administrator and the other admin roles, Target resources = All cloud apps.
  2. Session → Sign-in frequency = 4 hours (or less), and Persistent browser session = Never persistent.
  3. Turn it On. Admins sign in again every few hours, and closing the browser signs them out.

Maps to:

  • AC.L2-3.1.11 Terminate user sessions after a defined condition. Short admin sessions that end when the browser closes limit how long a stolen session or an unattended computer can be used with admin rights.
  • 164.312(a)(2)(iii) Automatic logoff. Short admin sessions that end when the browser closes limit how long a stolen session or an unattended computer can be used with admin rights.
  • 4.3 Configure Automatic Session Locking on Enterprise Assets. Short admin sessions that end when the browser closes limit how long a stolen session or an unattended computer can be used with admin rights.
  • 314.4(c)(1) Access controls. Short admin sessions that end when the browser closes limit how long a stolen session or an unattended computer can be used with admin rights.
  • PR.AA-03 Users authenticated. Short admin sessions that end when the browser closes limit how long a stolen session or an unattended computer can be used with admin rights.
  • 8.2.8 Idle sessions re-authenticate. Short admin sessions that end when the browser closes limit how long a stolen session or an unattended computer can be used with admin rights.
  • CC6.3 Least privilege. Short admin sessions that end when the browser closes limit how long a stolen session or an unattended computer can be used with admin rights.
  • CI-12 Account and password hygiene. Short admin sessions that end when the browser closes limit how long a stolen session or an unattended computer can be used with admin rights.

Trimtab's own access is read-only (M365-026)

High Privileged access

Fix: Remove any admin role (active or eligible), extra API or delegated permission, owned object, Azure role, Exchange role, or credential on the Trimtab M365 Scanner app in this tenant, leaving only Global Reader and the permissions Trimtab asks for. Then check the Entra audit log for who added it.

Take the helm: step-by-step
  1. Open the evidence to see what's extra: roles, API permissions, Exchange roles, or credentials.
  2. Roles: Entra admin center → Identity → Roles & admins → open each role named in the evidence → remove Trimtab M365 Scanner. Keep Global Reader. For a role marked eligible, remove it under Privileged Identity Management → Microsoft Entra roles → Assignments → Eligible assignments.
  3. API permissions or credentials: Entra admin center → Enterprise applications → Trimtab M365 Scanner → Permissions, and Certificates & secrets (or Properties). Remove anything Trimtab doesn't list in its documentation. If you can't remove a permission there, delete the app and connect Microsoft 365 again.
  4. Delegated permissions: Enterprise applications → Trimtab M365 Scanner → Permissions → User consent and Admin consent tabs; revoke anything beyond 'Sign you in' (openid). Owned objects: open each app named in the evidence → Owners → remove Trimtab M365 Scanner. Azure: Azure portal → each subscription or management group named → Access control (IAM) → Role assignments → remove Trimtab M365 Scanner.
  5. Exchange roles: in Exchange Online PowerShell, Get-ManagementRoleAssignment -RoleAssignee <Trimtab service principal>, then Remove-ManagementRoleAssignment for each one.
  6. Then check who added it: Entra admin center → Monitoring & health → Audit logs, filtered to Trimtab M365 Scanner. If you didn't add it, treat the tenant as compromised and tell Trimtab.
  7. If the check warns instead, Trimtab couldn't read one part (named in the result). For Exchange, assign Global Reader to Trimtab M365 Scanner; otherwise select Scan now later.

Maps to:

  • AC.L2-3.1.5 Employ the principle of least privilege. Trimtab only needs to read. Extra roles, permissions, or credentials on its app are access nobody uses day to day, and adding them is how attackers keep a foothold through a vendor's app.
  • AC.L1-B.1.II Limit access to permitted functions. Trimtab only needs to read. Extra roles, permissions, or credentials on its app are access nobody uses day to day, and adding them is how attackers keep a foothold through a vendor's app.
  • 164.312(a)(1) Access control. Trimtab only needs to read. Extra roles, permissions, or credentials on its app are access nobody uses day to day, and adding them is how attackers keep a foothold through a vendor's app.
  • 314.4(c)(1) Access controls. Trimtab only needs to read. Extra roles, permissions, or credentials on its app are access nobody uses day to day, and adding them is how attackers keep a foothold through a vendor's app.
  • PR.AA-05 Least privilege. Trimtab only needs to read. Extra roles, permissions, or credentials on its app are access nobody uses day to day, and adding them is how attackers keep a foothold through a vendor's app.
  • 7.2.2 Least privilege. Trimtab only needs to read. Extra roles, permissions, or credentials on its app are access nobody uses day to day, and adding them is how attackers keep a foothold through a vendor's app.
  • 500.7(a)(2) Limit privileged accounts. Trimtab only needs to read. Extra roles, permissions, or credentials on its app are access nobody uses day to day, and adding them is how attackers keep a foothold through a vendor's app.
  • 248.30(a)(2)(iii) Protection against unauthorized access. Trimtab only needs to read. Extra roles, permissions, or credentials on its app are access nobody uses day to day, and adding them is how attackers keep a foothold through a vendor's app.
  • CC6.3 Least privilege. Trimtab only needs to read. Extra roles, permissions, or credentials on its app are access nobody uses day to day, and adding them is how attackers keep a foothold through a vendor's app.

Applications

Users cannot consent to risky apps (M365-006)

Medium Applications

Fix: Set user consent to 'Do not allow' or 'Allow for verified publishers, for selected permissions', and use the admin consent workflow.

Take the helm: step-by-step
  1. Entra admin center → Enterprise applications → Consent and permissions → User consent settings → 'Do not allow user consent' (or 'Allow user consent for apps from verified publishers, for selected permissions').
  2. Turn on Admin consent requests (Enterprise applications → Admin consent settings) so users can ask for approval.

Maps to:

  • CM.L2-3.4.9 Control and monitor user-installed software. Stopping users from approving apps on their own keeps unknown apps from reading company data.
  • 164.312(a)(1) Access control. Apps a user approves can read their mail and files. Admin review keeps patient data away from unvetted apps.
  • AC.L1-B.1.I Limit access to authorized users. Only apps an admin has reviewed get to act on users' behalf.
  • 4.1 Establish and Maintain a Secure Configuration Process. Stopping users from approving apps on their own keeps unknown apps from reading company data.
  • 314.4(c)(1) Access controls. Stopping users from approving apps on their own keeps unknown apps from reading company data.
  • PR.AA-05 Least privilege. Stopping users from approving apps on their own keeps unknown apps from reading company data.
  • 7.2.2 Least privilege. Stopping users from approving apps on their own keeps unknown apps from reading company data.
  • 500.3(g) Systems and network security. Stopping users from approving apps on their own keeps unknown apps from reading company data.
  • 248.30(a)(2)(iii) Protection against unauthorized access. Stopping users from approving apps on their own keeps unknown apps from reading company data.
  • CC6.3 Least privilege. Stopping users from approving apps on their own keeps unknown apps from reading company data.
  • CC6.8 Preventing malicious software. Stopping users from approving apps on their own keeps unknown apps from reading company data.
  • CI-09 Hardened configuration. Stopping users from approving apps on their own keeps unknown apps from reading company data.

Users cannot register applications (M365-007)

Low Applications

Fix: Set 'Users can register applications' to No in Entra ID user settings.

Take the helm: step-by-step
  1. Entra admin center → Users → User settings → 'Users can register applications' = No.
  2. Give people who build apps the Application Developer role instead.

Maps to:

  • CM.L2-3.4.9 Control and monitor user-installed software. Blocking users from registering apps keeps app creation, and the permissions that come with it, in admin hands.
  • AC.L2-3.1.5 Employ the principle of least privilege. Blocking users from registering apps keeps app creation, and the permissions that come with it, in admin hands.
  • AC.L1-B.1.II Limit access to permitted functions. Blocking users from registering apps keeps app creation, and the permissions that come with it, in admin hands.
  • 4.1 Establish and Maintain a Secure Configuration Process. Blocking users from registering apps keeps app creation, and the permissions that come with it, in admin hands.
  • 314.4(c)(1) Access controls. Blocking users from registering apps keeps app creation, and the permissions that come with it, in admin hands.
  • PR.AA-05 Least privilege. Blocking users from registering apps keeps app creation, and the permissions that come with it, in admin hands.
  • 7.2.2 Least privilege. Blocking users from registering apps keeps app creation, and the permissions that come with it, in admin hands.
  • 500.7(a)(2) Limit privileged accounts. Blocking users from registering apps keeps app creation, and the permissions that come with it, in admin hands.
  • 248.30(a)(2)(iii) Protection against unauthorized access. Blocking users from registering apps keeps app creation, and the permissions that come with it, in admin hands.
  • CC6.3 Least privilege. Blocking users from registering apps keeps app creation, and the permissions that come with it, in admin hands.
  • CI-09 Hardened configuration. Blocking users from registering apps keeps app creation, and the permissions that come with it, in admin hands.

Admin consent requests are turned on (M365-017)

Low Applications

Fix: Turn on admin consent requests in Entra (Enterprise applications, Admin consent settings) and choose who reviews them.

Take the helm: step-by-step
  1. Entra admin center → Identity → Applications → Enterprise applications → Consent and permissions → Admin consent settings.
  2. Set 'Users can request admin consent to apps they are unable to consent to' = Yes, and choose the reviewers.
  3. Reviewers get an email for each request and approve or deny it in the Entra admin center.

Maps to:

  • CM.L2-3.4.9 Control and monitor user-installed software. With admin consent requests, users who need an app ask for it instead of being blocked or finding a workaround, and an admin reviews what it can access.
  • 314.4(c)(1) Access controls. With admin consent requests, users who need an app ask for it instead of being blocked or finding a workaround, and an admin reviews what it can access.
  • PR.PS-05 Unauthorized software prevented. With admin consent requests, users who need an app ask for it instead of being blocked or finding a workaround, and an admin reviews what it can access.
  • CC6.8 Preventing malicious software. With admin consent requests, users who need an app ask for it instead of being blocked or finding a workaround, and an admin reviews what it can access.

Users can't create new tenants (M365-018)

Low Applications

Fix: In Entra user settings, set 'Restrict non-admin users from creating tenants' to Yes.

Take the helm: step-by-step
  1. Entra admin center → Identity → Users → User settings.
  2. Set 'Restrict non-admin users from creating tenants' = Yes, then Save.

Maps to:

  • AC.L2-3.1.7 Prevent non-privileged users from executing privileged functions. A new tenant created by a user sits outside your control and settings. Restricting it keeps company work in the tenant you manage.
  • 314.4(c)(1) Access controls. A new tenant created by a user sits outside your control and settings. Restricting it keeps company work in the tenant you manage.
  • PR.AA-05 Least privilege. A new tenant created by a user sits outside your control and settings. Restricting it keeps company work in the tenant you manage.
  • CC6.3 Least privilege. A new tenant created by a user sits outside your control and settings. Restricting it keeps company work in the tenant you manage.

Users can't install Outlook add-ins (EXO-017)

Low Applications

Fix: Remove the My Custom Apps, My Marketplace Apps, and My ReadWriteMailbox Apps roles from the default role assignment policy, and deploy approved add-ins centrally.

Take the helm: step-by-step
  1. Exchange admin center → Roles → User roles → Default Role Assignment Policy.
  2. Clear My Custom Apps, My Marketplace Apps, and My ReadWriteMailbox Apps, then Save.
  3. Deploy the add-ins your users need centrally: Microsoft 365 admin center → Settings → Integrated apps.

Maps to:

  • CM.L2-3.4.9 Control and monitor user-installed software. Outlook add-ins can read and send email. Approving them centrally keeps unknown add-ins away from company mail.
  • 314.4(c)(1) Access controls. Outlook add-ins can read and send email. Approving them centrally keeps unknown add-ins away from company mail.
  • PR.PS-05 Unauthorized software prevented. Outlook add-ins can read and send email. Approving them centrally keeps unknown add-ins away from company mail.
  • CC6.8 Preventing malicious software. Outlook add-ins can read and send email. Approving them centrally keeps unknown add-ins away from company mail.
  • CI-09 Hardened configuration. Outlook add-ins can read and send email. Approving them centrally keeps unknown add-ins away from company mail.

External access

Guest access and invitations are restricted (M365-008)

Medium External access

Fix: Restrict guest user access to their own directory objects, and limit invitations to admins and the Guest Inviter role.

Take the helm: step-by-step
  1. Entra admin center → External Identities → External collaboration settings.
  2. Guest user access: 'Guest user access is restricted to properties and memberships of their own directory objects'.
  3. Guest invite settings: 'Only users assigned to specific admin roles can invite guest users'.

Maps to:

  • AC.L2-3.1.20 Verify and control connections to external systems. Limiting guest access, and who can invite guests, keeps outsiders out unless you approve them.
  • AC.L2-3.1.1 Limit system access to authorized users. Guests only get in when an approved person invites them.
  • 164.312(a)(1) Access control. Limiting guest access, and who can invite guests, keeps outsiders out unless you approve them.
  • AC.L1-B.1.I Limit access to authorized users. Limiting guest access, and who can invite guests, keeps outsiders out unless you approve them.
  • AC.L1-B.1.III Control external connections. Limiting guest access, and who can invite guests, keeps outsiders out unless you approve them.
  • 3.3 Configure Data Access Control Lists. Limiting guest access, and who can invite guests, keeps outsiders out unless you approve them.
  • 314.4(c)(1) Access controls. Limiting guest access, and who can invite guests, keeps outsiders out unless you approve them.
  • PR.AA-05 Least privilege. Limiting guest access, and who can invite guests, keeps outsiders out unless you approve them.
  • 7.2.2 Least privilege. Limiting guest access, and who can invite guests, keeps outsiders out unless you approve them.
  • 500.3(g) Systems and network security. Limiting guest access, and who can invite guests, keeps outsiders out unless you approve them.
  • 248.30(a)(2)(iii) Protection against unauthorized access. Limiting guest access, and who can invite guests, keeps outsiders out unless you approve them.
  • CC6.6 Protection from outside threats. Limiting guest access, and who can invite guests, keeps outsiders out unless you approve them.
  • CI-12 Account and password hygiene. Limiting guest access, and who can invite guests, keeps outsiders out unless you approve them.

SharePoint and OneDrive do not allow anonymous links (M365-009)

High External access

Fix: Set SharePoint external sharing to 'New and existing guests' or more restrictive, so 'Anyone' links are not allowed.

Take the helm: step-by-step
  1. SharePoint admin center → Policies → Sharing.
  2. Set the SharePoint and OneDrive slider to 'New and existing guests' or more restrictive, so 'Anyone' links aren't allowed.
  3. Optionally set the default link type to 'Specific people' and limit sharing to approved domains.

Maps to:

  • AC.L2-3.1.22 Control CUI on publicly accessible systems. 'Anyone with the link' shares open for anyone who has the URL, with no sign-in. Turning them off keeps files from going public.
  • AC.L2-3.1.20 Verify and control connections to external systems. 'Anyone with the link' shares open for anyone who has the URL, with no sign-in. Turning them off keeps files from going public.
  • 164.312(a)(1) Access control. Without anonymous links, every share needs a sign-in, so only permitted people reach patient data.
  • AC.L1-B.1.III Control external connections. 'Anyone with the link' shares open for anyone who has the URL, with no sign-in. Turning them off keeps files from going public.
  • AC.L1-B.1.IV Control public information. 'Anyone with the link' shares open for anyone who has the URL, with no sign-in. Turning them off keeps files from going public.
  • 3.3 Configure Data Access Control Lists. 'Anyone with the link' shares open for anyone who has the URL, with no sign-in. Turning them off keeps files from going public.
  • 314.4(c)(1) Access controls. 'Anyone with the link' shares open for anyone who has the URL, with no sign-in. Turning them off keeps files from going public.
  • PR.AA-05 Least privilege. 'Anyone with the link' shares open for anyone who has the URL, with no sign-in. Turning them off keeps files from going public.
  • PR.DS-01 Data at rest protected. 'Anyone with the link' shares open for anyone who has the URL, with no sign-in. Turning them off keeps files from going public.
  • 500.3(g) Systems and network security. 'Anyone with the link' shares open for anyone who has the URL, with no sign-in. Turning them off keeps files from going public.
  • 248.30(a)(2)(iii) Protection against unauthorized access. 'Anyone with the link' shares open for anyone who has the URL, with no sign-in. Turning them off keeps files from going public.
  • CC6.7 Controlling how data leaves. 'Anyone with the link' shares open for anyone who has the URL, with no sign-in. Turning them off keeps files from going public.
  • CI-12 Account and password hygiene. 'Anyone with the link' shares open for anyone who has the URL, with no sign-in. Turning them off keeps files from going public.

No stale guest accounts (M365-019)

Low External access

Fix: Remove guests who haven't signed in for 90 days, and set up a regular guest access review.

Default thresholds: Inactivity period (days) = 90

Take the helm: step-by-step
  1. Review the guests listed in the evidence and confirm with whoever invited them.
  2. Delete guests who no longer need access: Entra admin center → Users → filter User type = Guest → select → Delete.
  3. With Entra ID P2, set up a recurring access review for guests (Identity Governance → Access reviews).

Maps to:

  • AC.L2-3.1.1 Limit system access to authorized users. Guests who no longer work with you still have access to whatever was shared with them. Removing stale guests closes doors nobody is watching.
  • IA.L2-3.5.6 Disable identifiers after a period of inactivity. Guest accounts unused for 90 days are found so they can be disabled or removed.
  • 164.312(a)(1) Access control. Guests who no longer work with you still have access to whatever was shared with them. Removing stale guests closes doors nobody is watching.
  • 5.3 Disable Dormant Accounts. Guests who no longer work with you still have access to whatever was shared with them. Removing stale guests closes doors nobody is watching.
  • 314.4(c)(1) Access controls. Guests who no longer work with you still have access to whatever was shared with them. Removing stale guests closes doors nobody is watching.
  • PR.AA-01 Identities and credentials managed. Guests who no longer work with you still have access to whatever was shared with them. Removing stale guests closes doors nobody is watching.
  • 8.2.6 Inactive accounts removed. Guests who no longer work with you still have access to whatever was shared with them. Removing stale guests closes doors nobody is watching.
  • 500.7(a)(4) Remove unneeded accounts. Guests who no longer work with you still have access to whatever was shared with them. Removing stale guests closes doors nobody is watching.
  • 248.30(a)(2)(iii) Protection against unauthorized access. Guests who no longer work with you still have access to whatever was shared with them. Removing stale guests closes doors nobody is watching.
  • CC6.2 Granting and removing access. Guests who no longer work with you still have access to whatever was shared with them. Removing stale guests closes doors nobody is watching.
  • CI-12 Account and password hygiene. Guests who no longer work with you still have access to whatever was shared with them. Removing stale guests closes doors nobody is watching.

SharePoint blocks legacy authentication (SPO-001)

Medium External access

Fix: In the SharePoint admin center, turn off apps that don't use modern authentication (Policies, Access control).

Take the helm: step-by-step
  1. SharePoint admin center → Policies → Access control → Apps that don't use modern authentication → Block access.
  2. Or run Set-SPOTenant -LegacyAuthProtocolsEnabled $false in SharePoint Online PowerShell.

Maps to:

  • CM.L2-3.4.7 Restrict nonessential programs, functions, ports, protocols, and services. Legacy authentication is an outdated protocol that should be turned off.
  • IA.L2-3.5.3 Use multifactor authentication. Old SharePoint sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.
  • IA.L1-B.1.VI Authenticate users and devices. Old SharePoint sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.
  • 164.312(d) Person or entity authentication. Old SharePoint sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.
  • 6.3 Require MFA for Externally-Exposed Applications. Old SharePoint sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.
  • 314.4(c)(5) Multifactor authentication. Old SharePoint sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.
  • PR.AA-03 Users authenticated. Old SharePoint sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.
  • 2.2.5 Insecure services justified and secured. Old SharePoint sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.
  • 8.4.2 MFA for access into the CDE. Old SharePoint sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.
  • 500.12(a) Multi-factor authentication. Old SharePoint sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.
  • 248.30(a)(2)(iii) Protection against unauthorized access. Old SharePoint sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.
  • CC6.6 Protection from outside threats. Old SharePoint sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.
  • CI-01 MFA for email and cloud apps. Old SharePoint sign-in methods can't do MFA, so attackers use them to get around it. Blocking them closes that gap.

Guests can't reshare files (SPO-002)

Low External access

Fix: In the SharePoint admin center (Policies, Sharing), clear 'Allow guests to share items they don't own'.

Take the helm: step-by-step
  1. SharePoint admin center → Policies → Sharing → More external sharing settings.
  2. Clear 'Allow guests to share items they don't own', then Save.

Maps to:

  • AC.L2-3.1.3 Control the flow of CUI. When guests can't reshare, files you share with one outside person don't spread further without your say.
  • AC.L2-3.1.22 Control CUI on publicly accessible systems. When guests can't reshare, files you share with one outside person don't spread further without your say.
  • AC.L1-B.1.IV Control public information. When guests can't reshare, files you share with one outside person don't spread further without your say.
  • 164.312(a)(1) Access control. When guests can't reshare, files you share with one outside person don't spread further without your say.
  • 3.3 Configure Data Access Control Lists. When guests can't reshare, files you share with one outside person don't spread further without your say.
  • 314.4(c)(1) Access controls. When guests can't reshare, files you share with one outside person don't spread further without your say.
  • PR.AA-05 Least privilege. When guests can't reshare, files you share with one outside person don't spread further without your say.
  • 248.30(a)(2)(iii) Protection against unauthorized access. When guests can't reshare, files you share with one outside person don't spread further without your say.
  • CC6.7 Controlling how data leaves. When guests can't reshare, files you share with one outside person don't spread further without your say.

OneDrive syncs only on managed devices (SPO-003)

Low External access

Fix: In the SharePoint admin center (Settings, OneDrive Sync), allow syncing only on computers joined to your domains.

Take the helm: step-by-step
  1. Find your domain GUIDs: Get-ADDomain | Select ObjectGUID on a domain controller.
  2. SharePoint admin center → Settings → OneDrive - Sync → 'Allow syncing only on computers joined to specific domains', and add the GUIDs.
  3. For Entra-joined computers without on-premises AD, use Conditional Access app controls or Intune instead.

Maps to:

  • AC.L2-3.1.3 Control the flow of CUI. Limiting OneDrive sync to managed computers keeps copies of company files off personal and unknown devices.
  • 164.312(a)(1) Access control. Limiting OneDrive sync to managed computers keeps copies of company files off personal and unknown devices.
  • 314.4(c)(2) Know your systems and devices. Limiting OneDrive sync to managed computers keeps copies of company files off personal and unknown devices.
  • PR.DS-01 Data at rest protected. Limiting OneDrive sync to managed computers keeps copies of company files off personal and unknown devices.
  • 500.3(c) Asset inventory, device management, and end of life. Limiting OneDrive sync to managed computers keeps copies of company files off personal and unknown devices.
  • 248.30(a)(2)(iii) Protection against unauthorized access. Limiting OneDrive sync to managed computers keeps copies of company files off personal and unknown devices.
  • CC6.7 Controlling how data leaves. Limiting OneDrive sync to managed computers keeps copies of company files off personal and unknown devices.

SharePoint sharing is limited to approved domains (SPO-004)

Low External access

Fix: In the SharePoint admin center (Policies, Sharing, More external sharing settings), limit external sharing to specific domains: the partners you work with.

Take the helm: step-by-step
  1. List the outside organizations your people share files with.
  2. SharePoint admin center → Policies → Sharing → More external sharing settings → 'Limit external sharing by domain' → Add domains → 'Allow only specific domains', and add them.
  3. Add new partners as they come up. Sharing with other domains is refused.

Maps to:

  • AC.L2-3.1.20 Verify and control connections to external systems. Only approved outside organizations can receive shared files, which controls connections to external systems.
  • AC.L2-3.1.3 Control the flow of CUI. Limiting sharing to approved partner domains means files can only go to organizations you've chosen, not to any address someone types.
  • AC.L1-B.1.III Control external connections. Only approved outside organizations can receive shared files, which controls connections to external systems.
  • 164.312(a)(1) Access control. Limiting sharing to approved partner domains means files can only go to organizations you've chosen, not to any address someone types.
  • 314.4(c)(1) Access controls. Limiting sharing to approved partner domains means files can only go to organizations you've chosen, not to any address someone types.
  • PR.AA-05 Least privilege. Limiting sharing to approved partner domains means files can only go to organizations you've chosen, not to any address someone types.
  • 248.30(a)(2)(iii) Protection against unauthorized access. Limiting sharing to approved partner domains means files can only go to organizations you've chosen, not to any address someone types.
  • CC6.6 Protection from outside threats. Limiting sharing to approved partner domains means files can only go to organizations you've chosen, not to any address someone types.

Calendar details aren't shared outside the company (EXO-018)

Low External access

Fix: Limit the default sharing policy to free/busy time only (or turn external calendar sharing off), so meeting subjects and attendees stay inside.

Take the helm: step-by-step
  1. Exchange admin center → Organization → Sharing → Individual sharing → open the Default Sharing Policy.
  2. Set the rule for all domains to 'Calendar sharing with free/busy information only', or clear the policy to turn external calendar sharing off.
  3. Or in PowerShell: Set-SharingPolicy -Identity 'Default Sharing Policy' -Domains '*:CalendarSharingFreeBusySimple'.

Maps to:

  • AC.L2-3.1.3 Control the flow of CUI. Detailed calendar sharing shows meeting subjects, attendees, and locations to outsiders. Free/busy only shares when people are available.
  • 164.312(a)(1) Access control. Detailed calendar sharing shows meeting subjects, attendees, and locations to outsiders. Free/busy only shares when people are available.
  • 314.4(c)(1) Access controls. Detailed calendar sharing shows meeting subjects, attendees, and locations to outsiders. Free/busy only shares when people are available.
  • PR.DS-01 Data at rest protected. Detailed calendar sharing shows meeting subjects, attendees, and locations to outsiders. Free/busy only shares when people are available.
  • 248.30(a)(2)(i) Security and confidentiality of customer information. Detailed calendar sharing shows meeting subjects, attendees, and locations to outsiders. Free/busy only shares when people are available.
  • CC6.7 Controlling how data leaves. Detailed calendar sharing shows meeting subjects, attendees, and locations to outsiders. Free/busy only shares when people are available.

Posture

Microsoft Secure Score meets target (M365-011)

Low Posture

Fix: Work through the improvement actions in Microsoft Secure Score, starting with the highest impact.

Default thresholds: Target percentage = 50

Take the helm: step-by-step
  1. Open the Microsoft Defender portal → Exposure management → Secure Score → Recommended actions.
  2. Sort by score impact and work through the top items. Many are covered by other Trimtab checks.
  3. Mark items handled by third-party tools as resolved that way, so the score reflects reality.

Maps to:

  • CM.L2-3.4.2 Establish and enforce security configuration settings. Secure Score measures how many of Microsoft's recommended security settings are in place.
  • 164.308(a)(1)(ii)(B) Risk management. Secure Score lists your weak spots in priority order, which feeds straight into risk management.
  • 4.1 Establish and Maintain a Secure Configuration Process. Secure Score measures how many of Microsoft's recommended security settings are in place.
  • 314.4(d)(1) Test and monitor safeguards. Tracking Secure Score is a regular, measurable check that your safeguards are working.
  • ID.RA-01 Vulnerabilities identified. Secure Score measures how many of Microsoft's recommended security settings are in place.
  • 500.3(g) Systems and network security. Secure Score measures how many of Microsoft's recommended security settings are in place.
  • 248.30(a)(2)(ii) Protection against anticipated threats. Secure Score measures how many of Microsoft's recommended security settings are in place.
  • CC7.1 Finding vulnerabilities and configuration drift. Secure Score measures how many of Microsoft's recommended security settings are in place.

Exchange Online

Mailbox auditing is on for the organization (EXO-001)

High Exchange Online

Fix: Run Set-OrganizationConfig -AuditDisabled $false so mailbox actions are logged for every mailbox.

Take the helm: step-by-step
  1. Connect to Exchange Online PowerShell: Connect-ExchangeOnline.
  2. Run Set-OrganizationConfig -AuditDisabled $false.
  3. Confirm with Get-OrganizationConfig | Format-List AuditDisabled.

Maps to:

  • AU.L2-3.3.1 Create and retain system audit logs. Mailbox auditing records who opened, sent, or deleted email in each mailbox.
  • AU.L2-3.3.2 Ensure actions of individual users can be traced. Mailbox auditing records who opened, sent, or deleted email in each mailbox.
  • 164.312(b) Audit controls. Mailbox auditing records who opened, sent, or deleted email in each mailbox.
  • 8.2 Collect Audit Logs. Mailbox auditing records who opened, sent, or deleted email in each mailbox.
  • 314.4(c)(8) Monitor and log user activity. Mailbox auditing records who opened, sent, or deleted email in each mailbox.
  • PR.PS-04 Log records generated. Mailbox auditing records who opened, sent, or deleted email in each mailbox.
  • DE.CM-03 Personnel activity monitored. Mailbox auditing records who opened, sent, or deleted email in each mailbox.
  • 10.2.1 Audit logs enabled. Mailbox auditing records who opened, sent, or deleted email in each mailbox.
  • 500.6(a)(2) Audit trails. Mailbox auditing records who opened, sent, or deleted email in each mailbox.
  • 248.30(a)(2)(i) Security and confidentiality of customer information. Mailbox auditing records who opened, sent, or deleted email in each mailbox.
  • CC7.2 Monitoring for attacks. Mailbox auditing records who opened, sent, or deleted email in each mailbox.
  • CI-11 Logging. Mailbox auditing records who opened, sent, or deleted email in each mailbox.

Unified audit log is on (EXO-002)

High Exchange Online

Fix: Turn on auditing in the Microsoft Purview portal, or run Set-AdminAuditLogConfig -UnifiedAuditLogIngestionEnabled $true.

Take the helm: step-by-step
  1. Microsoft Purview portal → Audit → 'Start recording user and admin activity'.
  2. Or in Exchange Online PowerShell: Set-AdminAuditLogConfig -UnifiedAuditLogIngestionEnabled $true.
  3. Recording can take up to an hour to start.

Maps to:

  • AU.L2-3.3.1 Create and retain system audit logs. The unified audit log records user and admin activity across Microsoft 365. It's the main trail for any investigation.
  • AU.L2-3.3.2 Ensure actions of individual users can be traced. The unified audit log records user and admin activity across Microsoft 365. It's the main trail for any investigation.
  • 164.312(b) Audit controls. The unified audit log records user and admin activity across Microsoft 365. It's the main trail for any investigation.
  • 8.2 Collect Audit Logs. The unified audit log records user and admin activity across Microsoft 365. It's the main trail for any investigation.
  • 314.4(c)(8) Monitor and log user activity. The unified audit log records user and admin activity across Microsoft 365. It's the main trail for any investigation.
  • PR.PS-04 Log records generated. The unified audit log records user and admin activity across Microsoft 365. It's the main trail for any investigation.
  • DE.CM-03 Personnel activity monitored. The unified audit log records user and admin activity across Microsoft 365. It's the main trail for any investigation.
  • 10.2.1 Audit logs enabled. The unified audit log records user and admin activity across Microsoft 365. It's the main trail for any investigation.
  • 500.6(a)(2) Audit trails. The unified audit log records user and admin activity across Microsoft 365. It's the main trail for any investigation.
  • 248.30(a)(2)(i) Security and confidentiality of customer information. The unified audit log records user and admin activity across Microsoft 365. It's the main trail for any investigation.
  • CC7.2 Monitoring for attacks. The unified audit log records user and admin activity across Microsoft 365. It's the main trail for any investigation.
  • CI-11 Logging. The unified audit log records user and admin activity across Microsoft 365. It's the main trail for any investigation.

Automatic forwarding to external addresses is blocked (EXO-003)

High Exchange Online

Fix: Set automatic forwarding to Off in the outbound spam filter policy, and allow it only for approved users with a separate policy.

Take the helm: step-by-step
  1. Microsoft Defender portal → Email & collaboration → Policies & rules → Threat policies → Anti-spam → Anti-spam outbound policy (Default).
  2. Set Automatic forwarding rules to 'Off - Forwarding is disabled'.
  3. If a few users truly need forwarding, create a separate outbound policy just for them.
  4. PowerShell: Set-HostedOutboundSpamFilterPolicy -Identity Default -AutoForwardingMode Off.

Maps to:

  • AC.L2-3.1.3 Control the flow of CUI. Blocking automatic external forwarding stops email, and whatever is in it, from quietly flowing to outside addresses.
  • 164.312(e)(1) Transmission security. Blocking automatic external forwarding stops email, and whatever is in it, from quietly flowing to outside addresses.
  • AC.L1-B.1.III Control external connections. Blocking automatic external forwarding stops email, and whatever is in it, from quietly flowing to outside addresses.
  • SC.L1-B.1.X Protect communications at the boundary. The policy controls what email can leave at the edge of your mail system.
  • 3.3 Configure Data Access Control Lists. Blocking automatic external forwarding stops email, and whatever is in it, from quietly flowing to outside addresses.
  • 314.4(c)(1) Access controls. Blocking automatic external forwarding stops email, and whatever is in it, from quietly flowing to outside addresses.
  • PR.DS-02 Data in transit protected. Blocking automatic external forwarding stops email, and whatever is in it, from quietly flowing to outside addresses.
  • 500.3(g) Systems and network security. Blocking automatic external forwarding stops email, and whatever is in it, from quietly flowing to outside addresses.
  • 248.30(a)(2)(i) Security and confidentiality of customer information. Blocking automatic external forwarding stops email, and whatever is in it, from quietly flowing to outside addresses.
  • CC6.7 Controlling how data leaves. Blocking automatic external forwarding stops email, and whatever is in it, from quietly flowing to outside addresses.
  • CI-07 Email security. Blocking automatic external forwarding stops email, and whatever is in it, from quietly flowing to outside addresses.

No mailboxes forward to external addresses (EXO-004)

Medium Exchange Online

Fix: Remove external forwarding addresses from mailboxes unless there is a documented business need.

Take the helm: step-by-step
  1. Review each mailbox in the evidence with its owner.
  2. Remove the forward: Exchange admin center → Recipients → Mailboxes → mailbox → Mail flow settings → Email forwarding, or Set-Mailbox <user> -ForwardingSmtpAddress $null -DeliverToMailboxAndForward $false.
  3. Check the user's inbox rules too: Get-InboxRule -Mailbox <user>.

Maps to:

  • AC.L2-3.1.3 Control the flow of CUI. A mailbox forwarding to an outside address sends every message out of the company. This finds any that do.
  • 164.312(e)(1) Transmission security. A mailbox forwarding to an outside address sends every message out of the company. This finds any that do.
  • AC.L1-B.1.III Control external connections. A mailbox forwarding to an outside address sends every message out of the company. This finds any that do.
  • SC.L1-B.1.X Protect communications at the boundary. Finding forwarding mailboxes shows what email is leaving at the edge of your mail system.
  • 3.3 Configure Data Access Control Lists. A mailbox forwarding to an outside address sends every message out of the company. This finds any that do.
  • 314.4(c)(1) Access controls. A mailbox forwarding to an outside address sends every message out of the company. This finds any that do.
  • PR.DS-02 Data in transit protected. A mailbox forwarding to an outside address sends every message out of the company. This finds any that do.
  • 500.3(g) Systems and network security. A mailbox forwarding to an outside address sends every message out of the company. This finds any that do.
  • 248.30(a)(2)(i) Security and confidentiality of customer information. A mailbox forwarding to an outside address sends every message out of the company. This finds any that do.
  • CC6.7 Controlling how data leaves. A mailbox forwarding to an outside address sends every message out of the company. This finds any that do.
  • CI-07 Email security. A mailbox forwarding to an outside address sends every message out of the company. This finds any that do.

Modern authentication is on for Exchange (EXO-005)

Medium Exchange Online

Fix: Run Set-OrganizationConfig -OAuth2ClientProfileEnabled $true so Outlook uses modern authentication and MFA.

Take the helm: step-by-step
  1. In Exchange Online PowerShell, run Set-OrganizationConfig -OAuth2ClientProfileEnabled $true.
  2. Newer tenants have this on by default; older tenants may need it turned on.

Maps to:

  • IA.L2-3.5.3 Use multifactor authentication. Modern authentication lets Outlook use MFA. Without it, email apps sign in with just a password.
  • 164.312(d) Person or entity authentication. Modern authentication lets Outlook use MFA. Without it, email apps sign in with just a password.
  • IA.L1-B.1.VI Authenticate users and devices. Modern authentication lets Outlook use MFA. Without it, email apps sign in with just a password.
  • 6.3 Require MFA for Externally-Exposed Applications. Modern authentication lets Outlook use MFA. Without it, email apps sign in with just a password.
  • 314.4(c)(5) Multifactor authentication. Modern authentication lets Outlook use MFA. Without it, email apps sign in with just a password.
  • PR.AA-03 Users authenticated. Modern authentication lets Outlook use MFA. Without it, email apps sign in with just a password.
  • 8.4.2 MFA for access into the CDE. Modern authentication lets Outlook use MFA. Without it, email apps sign in with just a password.
  • 500.12(a) Multi-factor authentication. Modern authentication lets Outlook use MFA. Without it, email apps sign in with just a password.
  • 248.30(a)(2)(iii) Protection against unauthorized access. Modern authentication lets Outlook use MFA. Without it, email apps sign in with just a password.
  • CC6.6 Protection from outside threats. Modern authentication lets Outlook use MFA. Without it, email apps sign in with just a password.
  • CI-01 MFA for email and cloud apps. Modern authentication lets Outlook use MFA. Without it, email apps sign in with just a password.

SMTP AUTH is turned off (EXO-006)

Medium Exchange Online

Fix: Run Set-TransportConfig -SmtpClientAuthenticationDisabled $true, and enable it only on mailboxes that need it, such as scanners.

Take the helm: step-by-step
  1. Find what still uses SMTP AUTH, such as scanners or line-of-business apps (sign-in logs or Exchange reports).
  2. Turn it off for the organization: Set-TransportConfig -SmtpClientAuthenticationDisabled $true.
  3. Re-enable it only on mailboxes that need it: Set-CASMailbox <mailbox> -SmtpClientAuthenticationDisabled $false.

Maps to:

  • CM.L2-3.4.7 Restrict nonessential programs, functions, ports, protocols, and services. SMTP AUTH is an older protocol most organizations no longer need.
  • IA.L2-3.5.3 Use multifactor authentication. SMTP AUTH sends mail with a password only, skipping MFA. Turning it off closes that back door.
  • 164.312(d) Person or entity authentication. SMTP AUTH sends mail with a password only, skipping MFA. Turning it off closes that back door.
  • IA.L1-B.1.VI Authenticate users and devices. SMTP AUTH sends mail with a password only, skipping MFA. Turning it off closes that back door.
  • 6.3 Require MFA for Externally-Exposed Applications. SMTP AUTH sends mail with a password only, skipping MFA. Turning it off closes that back door.
  • 314.4(c)(5) Multifactor authentication. SMTP AUTH sends mail with a password only, skipping MFA. Turning it off closes that back door.
  • PR.AA-03 Users authenticated. SMTP AUTH sends mail with a password only, skipping MFA. Turning it off closes that back door.
  • 2.2.5 Insecure services justified and secured. SMTP AUTH sends mail with a password only, skipping MFA. Turning it off closes that back door.
  • 8.4.2 MFA for access into the CDE. SMTP AUTH sends mail with a password only, skipping MFA. Turning it off closes that back door.
  • 500.12(a) Multi-factor authentication. SMTP AUTH sends mail with a password only, skipping MFA. Turning it off closes that back door.
  • 248.30(a)(2)(iii) Protection against unauthorized access. SMTP AUTH sends mail with a password only, skipping MFA. Turning it off closes that back door.
  • CC6.6 Protection from outside threats. SMTP AUTH sends mail with a password only, skipping MFA. Turning it off closes that back door.
  • CI-01 MFA for email and cloud apps. SMTP AUTH sends mail with a password only, skipping MFA. Turning it off closes that back door.

DKIM signing is on for custom domains (EXO-007)

Medium Exchange Online

Fix: Enable DKIM for each custom domain in the Microsoft Defender portal and publish the two CNAME records.

Take the helm: step-by-step
  1. Microsoft Defender portal → Email & collaboration → Policies & rules → Threat policies → Email authentication settings → DKIM.
  2. Select each domain and copy its two CNAME records (selector1._domainkey and selector2._domainkey).
  3. Add both records at your DNS host, wait for them to publish, then turn on 'Sign messages for this domain with DKIM signatures'.

Maps to:

  • SC.L2-3.13.1 Monitor, control, and protect communications at system boundaries. DKIM digitally signs your outgoing email, so other mail systems can tell it really came from you.
  • 164.308(a)(1)(ii)(B) Risk management. Fake email from your domain is a common way into organizations. DKIM is one of the protections that lowers that risk.
  • SC.L1-B.1.X Protect communications at the boundary. DKIM digitally signs your outgoing email, so other mail systems can tell it really came from you.
  • 314.4(c) Safeguards that control identified risks. DKIM digitally signs your outgoing email, so other mail systems can tell it really came from you.
  • PR.DS-02 Data in transit protected. DKIM signatures let receiving mail systems confirm your messages weren't forged or changed along the way.
  • 5.4.1 Protection against phishing. DKIM digitally signs your outgoing email, so other mail systems can tell it really came from you.
  • 500.14(a)(2) Protection against malicious code. DKIM digitally signs your outgoing email, so other mail systems can tell it really came from you.
  • 248.30(a)(2)(ii) Protection against anticipated threats. DKIM digitally signs your outgoing email, so other mail systems can tell it really came from you.
  • CC6.6 Protection from outside threats. DKIM digitally signs your outgoing email, so other mail systems can tell it really came from you.
  • CI-07 Email security. DKIM digitally signs your outgoing email, so other mail systems can tell it really came from you.

Malware filter blocks risky attachment types (EXO-009)

Medium Exchange Online

Fix: Turn on the common attachments filter in the default anti-malware policy.

Take the helm: step-by-step
  1. Microsoft Defender portal → Email & collaboration → Policies & rules → Threat policies → Anti-malware.
  2. Open the Default policy (and any custom policies) → Edit protection settings → turn on 'Enable the common attachments filter', and keep the default file types.
  3. Or in Exchange Online PowerShell: Set-MalwareFilterPolicy -Identity Default -EnableFileFilter $true.

Maps to:

  • SI.L2-3.14.2 Provide protection from malicious code. The attachment filter blocks file types commonly used to deliver malware by email.
  • 164.308(a)(5)(ii)(B) Protection from malicious software. The attachment filter blocks file types commonly used to deliver malware by email.
  • SI.L1-B.1.XIII Protect against malicious code. The attachment filter blocks file types commonly used to deliver malware by email.
  • 10.1 Deploy and Maintain Anti-Malware Software. The attachment filter blocks file types commonly used to deliver malware by email.
  • 314.4(c) Safeguards that control identified risks. The attachment filter blocks file types commonly used to deliver malware by email.
  • PR.PS-05 Unauthorized software prevented. The attachment filter blocks file types commonly used to deliver malware by email.
  • 5.4.1 Protection against phishing. The attachment filter blocks file types commonly used to deliver malware by email.
  • 500.14(a)(2) Protection against malicious code. The attachment filter blocks file types commonly used to deliver malware by email.
  • 248.30(a)(2)(ii) Protection against anticipated threats. The attachment filter blocks file types commonly used to deliver malware by email.
  • CC6.8 Preventing malicious software. The attachment filter blocks file types commonly used to deliver malware by email.
  • CI-07 Email security. The attachment filter blocks file types commonly used to deliver malware by email.

Shared mailboxes can't be signed in to (EXO-016)

Medium Exchange Online

Fix: Block sign-in for each shared mailbox's user account. People use shared mailboxes through their own accounts, with permissions.

Take the helm: step-by-step
  1. Microsoft 365 admin center → Users → Active users. Open each shared mailbox listed in the evidence.
  2. Select Block sign-in, then Save. The mailbox keeps working for everyone who has permission to it.
  3. Or in PowerShell: Update-MgUser -UserId <address> -AccountEnabled:$false.

Maps to:

  • AC.L2-3.1.1 Limit system access to authorized users. Shared mailboxes have no owner and often a forgotten password. Blocking sign-in means they can only be used through people's own accounts.
  • IA.L1-B.1.V Identify users and devices. Blocking direct sign-in means every use of the mailbox happens through a person's own account.
  • 164.312(a)(1) Access control. Shared mailboxes have no owner and often a forgotten password. Blocking sign-in means they can only be used through people's own accounts.
  • 314.4(c)(1) Access controls. Shared mailboxes have no owner and often a forgotten password. Blocking sign-in means they can only be used through people's own accounts.
  • PR.AA-01 Identities and credentials managed. Shared mailboxes have no owner and often a forgotten password. Blocking sign-in means they can only be used through people's own accounts.
  • 8.2.2 Shared accounts limited. With direct sign-in blocked, a shared mailbox can't be used as a shared account; every action traces to a person.
  • 248.30(a)(2)(iii) Protection against unauthorized access. Shared mailboxes have no owner and often a forgotten password. Blocking sign-in means they can only be used through people's own accounts.
  • CC6.1 Logical access security. Shared mailboxes have no owner and often a forgotten password. Blocking sign-in means they can only be used through people's own accounts.
  • CI-12 Account and password hygiene. Shared mailboxes have no owner and often a forgotten password. Blocking sign-in means they can only be used through people's own accounts.

Email security

DMARC is enforced for custom domains (EXO-008)

Medium Email security

Fix: Publish a DMARC record (_dmarc.yourdomain) with p=quarantine or p=reject after monitoring with p=none.

Take the helm: step-by-step
  1. Make sure SPF and DKIM are set up first.
  2. Add a TXT record at _dmarc.<yourdomain> with v=DMARC1; p=none; rua=mailto:reports@<yourdomain>.
  3. Review the reports for two to four weeks and fix any legitimate senders that fail.
  4. Change the policy to p=quarantine, and later p=reject.

Maps to:

  • SC.L2-3.13.1 Monitor, control, and protect communications at system boundaries. DMARC tells other mail systems to reject email that pretends to be from your domain.
  • 164.308(a)(1)(ii)(B) Risk management. Phishing that impersonates your domain is a top risk. Enforced DMARC lowers it.
  • SC.L1-B.1.X Protect communications at the boundary. DMARC tells other mail systems to reject email that pretends to be from your domain.
  • 314.4(c) Safeguards that control identified risks. DMARC tells other mail systems to reject email that pretends to be from your domain.
  • PR.DS-02 Data in transit protected. DMARC tells other mail systems to reject email that pretends to be from your domain.
  • 5.4.1 Protection against phishing. DMARC tells other mail systems to reject email that pretends to be from your domain.
  • 500.14(a)(2) Protection against malicious code. DMARC tells other mail systems to reject email that pretends to be from your domain.
  • 248.30(a)(2)(ii) Protection against anticipated threats. DMARC tells other mail systems to reject email that pretends to be from your domain.
  • CC6.6 Protection from outside threats. DMARC tells other mail systems to reject email that pretends to be from your domain.
  • CI-07 Email security. DMARC tells other mail systems to reject email that pretends to be from your domain.

SPF is published for custom domains (EXO-010)

Medium Email security

Fix: Publish one SPF record (a TXT record starting v=spf1) for each custom domain that lists the services allowed to send its mail and ends with ~all or -all.

Take the helm: step-by-step
  1. At your DNS host, look up the TXT records on the domain itself (not a subdomain). There must be exactly one that starts with v=spf1.
  2. For mail sent only through Microsoft 365, use v=spf1 include:spf.protection.outlook.com -all.
  3. Add an include: for each other service that sends as the domain (for example, a newsletter or billing system), keeping the whole record under 10 DNS lookups.
  4. End the record with ~all (soft fail) or -all (hard fail), never +all or ?all.
  5. For a domain that never sends email, publish v=spf1 -all so nobody can send as it.

Maps to:

  • SC.L2-3.13.1 Monitor, control, and protect communications at system boundaries. SPF lists the services allowed to send email as your domain, so other mail systems can spot forged messages.
  • 164.308(a)(1)(ii)(B) Risk management. Fake email from your domain is a common way into organizations. SPF is one of the protections that lowers that risk.
  • SC.L1-B.1.X Protect communications at the boundary. SPF lists the services allowed to send email as your domain, so other mail systems can spot forged messages.
  • 314.4(c) Safeguards that control identified risks. SPF lists the services allowed to send email as your domain, so other mail systems can spot forged messages.
  • PR.DS-02 Data in transit protected. SPF lists the services allowed to send email as your domain, so other mail systems can spot forged messages.
  • 5.4.1 Protection against phishing. SPF lists the services allowed to send email as your domain, so other mail systems can spot forged messages.
  • 500.14(a)(2) Protection against malicious code. SPF lists the services allowed to send email as your domain, so other mail systems can spot forged messages.
  • 248.30(a)(2)(ii) Protection against anticipated threats. SPF lists the services allowed to send email as your domain, so other mail systems can spot forged messages.
  • CC6.6 Protection from outside threats. SPF lists the services allowed to send email as your domain, so other mail systems can spot forged messages.
  • CI-07 Email security. SPF lists the services allowed to send email as your domain, so other mail systems can spot forged messages.

Anti-phishing protection is on (EXO-011)

Medium Email security

Fix: Keep spoof intelligence on in the anti-phishing policies. With Defender for Office 365, also turn on impersonation protection, for example with the Standard preset security policy.

Take the helm: step-by-step
  1. Microsoft Defender portal → Email & collaboration → Policies & rules → Threat policies → Anti-phishing.
  2. Open the default policy (Office365 AntiPhish Default) and any custom policies. Under Spoof, turn on spoof intelligence, 'Honor DMARC record policy', and 'Show (?) for unauthenticated senders'.
  3. With Defender for Office 365 (Business Premium includes it), the simplest way to turn on impersonation protection is Threat policies → Preset security policies → Standard protection, applied to all recipients.
  4. Or in a policy, turn on user impersonation (add executives and finance staff), domain impersonation for your own domains, and mailbox intelligence protection.
  5. Confirm in Exchange Online PowerShell: Get-AntiPhishPolicy | Format-List Name, EnableSpoofIntelligence, HonorDmarcPolicy, EnableMailboxIntelligenceProtection.

Maps to:

  • SI.L2-3.14.2 Provide protection from malicious code. Anti-phishing filters catch email that fakes a sender or impersonates your people, before it reaches inboxes.
  • 164.308(a)(5)(ii)(B) Protection from malicious software. Anti-phishing filters catch email that fakes a sender or impersonates your people, before it reaches inboxes.
  • SI.L1-B.1.XIII Protect against malicious code. Anti-phishing filters catch email that fakes a sender or impersonates your people, before it reaches inboxes.
  • 314.4(c) Safeguards that control identified risks. Anti-phishing filters catch email that fakes a sender or impersonates your people, before it reaches inboxes.
  • DE.CM-09 Computing monitored. Anti-phishing filters watch incoming email for signs of attack, such as spoofed or impersonated senders.
  • 5.4.1 Protection against phishing. Anti-phishing filters catch email that fakes a sender or impersonates your people, before it reaches inboxes.
  • 500.14(a)(2) Protection against malicious code. Anti-phishing filters catch email that fakes a sender or impersonates your people, before it reaches inboxes.
  • 248.30(a)(2)(ii) Protection against anticipated threats. Anti-phishing filters catch email that fakes a sender or impersonates your people, before it reaches inboxes.
  • CC6.8 Preventing malicious software. Anti-phishing filters catch email that fakes a sender or impersonates your people, before it reaches inboxes.
  • CI-07 Email security. Anti-phishing filters catch email that fakes a sender or impersonates your people, before it reaches inboxes.

Safe Links and Safe Attachments are on (EXO-012)

Medium Email security

Fix: With Defender for Office 365, turn on the Standard or Strict preset security policy (or Built-in protection), so links are checked when clicked and attachments are opened in a sandbox first.

Take the helm: step-by-step
  1. Needs Microsoft Defender for Office 365 (included in Microsoft 365 Business Premium).
  2. Microsoft Defender portal → Email & collaboration → Policies & rules → Threat policies → Preset security policies.
  3. Turn on Standard protection (or Strict) for all recipients. Built-in protection covers anyone not in a preset.

Maps to:

  • SI.L2-3.14.2 Provide protection from malicious code. Safe Links checks web links when someone clicks them, and Safe Attachments opens attachments in a sandbox first, catching threats ordinary filters miss.
  • SI.L2-3.14.5 Perform periodic and real-time scans. Every link and attachment is scanned in real time as email arrives and when it's opened.
  • SI.L1-B.1.XIII Protect against malicious code. Safe Links checks web links when someone clicks them, and Safe Attachments opens attachments in a sandbox first, catching threats ordinary filters miss.
  • SI.L1-B.1.XV Scan systems and files. Every link and attachment is scanned in real time as email arrives and when it's opened.
  • 164.308(a)(5)(ii)(B) Protection from malicious software. Safe Links checks web links when someone clicks them, and Safe Attachments opens attachments in a sandbox first, catching threats ordinary filters miss.
  • 314.4(c) Safeguards that control identified risks. Safe Links checks web links when someone clicks them, and Safe Attachments opens attachments in a sandbox first, catching threats ordinary filters miss.
  • DE.CM-09 Computing monitored. Safe Links checks web links when someone clicks them, and Safe Attachments opens attachments in a sandbox first, catching threats ordinary filters miss.
  • 5.2.1 Anti-malware deployed. Safe Links checks web links when someone clicks them, and Safe Attachments opens attachments in a sandbox first, catching threats ordinary filters miss.
  • 5.4.1 Protection against phishing. Safe Links checks web links when someone clicks them, and Safe Attachments opens attachments in a sandbox first, catching threats ordinary filters miss.
  • 500.14(a)(2) Protection against malicious code. Safe Links checks web links when someone clicks them, and Safe Attachments opens attachments in a sandbox first, catching threats ordinary filters miss.
  • 248.30(a)(2)(ii) Protection against anticipated threats. Safe Links checks web links when someone clicks them, and Safe Attachments opens attachments in a sandbox first, catching threats ordinary filters miss.
  • CC6.8 Preventing malicious software. Safe Links checks web links when someone clicks them, and Safe Attachments opens attachments in a sandbox first, catching threats ordinary filters miss.
  • CI-07 Email security. Safe Links checks web links when someone clicks them, and Safe Attachments opens attachments in a sandbox first, catching threats ordinary filters miss.
  • 10.1 Deploy and Maintain Anti-Malware Software. Safe Links checks web links when someone clicks them, and Safe Attachments opens attachments in a sandbox first, catching threats ordinary filters miss.

Mail flow rules don't skip spam filtering for outside senders (EXO-013)

Medium Email security

Fix: Remove or narrow mail flow rules that set the spam confidence level to -1 for sender domains or addresses. Use the Tenant Allow/Block List for specific, reviewed exceptions instead.

Take the helm: step-by-step
  1. Exchange admin center → Mail flow → Rules. Open each rule named in the evidence.
  2. Remove the 'Set the spam confidence level (SCL) to -1' action, or delete the rule.
  3. For a sender you've reviewed and trust, add an allow entry in the Tenant Allow/Block List, which still scans for malware and phishing.

Maps to:

  • SI.L2-3.14.2 Provide protection from malicious code. A mail flow rule that skips spam filtering for an outside domain lets anyone who fakes that domain straight into inboxes.
  • CM.L2-3.4.2 Establish and enforce security configuration settings. Mail flow rules that weaken filtering undo the secure email settings you've configured.
  • SI.L1-B.1.XIII Protect against malicious code. A mail flow rule that skips spam filtering for an outside domain lets anyone who fakes that domain straight into inboxes.
  • 164.308(a)(5)(ii)(B) Protection from malicious software. A mail flow rule that skips spam filtering for an outside domain lets anyone who fakes that domain straight into inboxes.
  • 314.4(c) Safeguards that control identified risks. A mail flow rule that skips spam filtering for an outside domain lets anyone who fakes that domain straight into inboxes.
  • DE.CM-09 Computing monitored. A mail flow rule that skips spam filtering for an outside domain lets anyone who fakes that domain straight into inboxes.
  • 5.4.1 Protection against phishing. A mail flow rule that skips spam filtering for an outside domain lets anyone who fakes that domain straight into inboxes.
  • 500.14(a)(2) Protection against malicious code. A mail flow rule that skips spam filtering for an outside domain lets anyone who fakes that domain straight into inboxes.
  • 248.30(a)(2)(ii) Protection against anticipated threats. A mail flow rule that skips spam filtering for an outside domain lets anyone who fakes that domain straight into inboxes.
  • CC6.8 Preventing malicious software. A mail flow rule that skips spam filtering for an outside domain lets anyone who fakes that domain straight into inboxes.
  • CI-07 Email security. A mail flow rule that skips spam filtering for an outside domain lets anyone who fakes that domain straight into inboxes.

Spam filters don't allow-list senders or IP addresses (EXO-014)

Medium Email security

Fix: Clear allowed sender domains and addresses in anti-spam policies and the IP allow list in the connection filter policy, and turn off the safe list.

Take the helm: step-by-step
  1. Microsoft Defender portal → Email & collaboration → Policies & rules → Threat policies → Anti-spam.
  2. In each inbound anti-spam policy, clear Allowed senders and Allowed domains.
  3. Open Connection filter policy (Default): clear the IP allow list and turn off 'Turn on safe list'.

Maps to:

  • SI.L2-3.14.2 Provide protection from malicious code. Allow lists skip spam filtering for listed senders and IP addresses, so a forged or compromised sender gets straight through.
  • SI.L1-B.1.XIII Protect against malicious code. Allow lists skip spam filtering for listed senders and IP addresses, so a forged or compromised sender gets straight through.
  • 164.308(a)(5)(ii)(B) Protection from malicious software. Allow lists skip spam filtering for listed senders and IP addresses, so a forged or compromised sender gets straight through.
  • 314.4(c) Safeguards that control identified risks. Allow lists skip spam filtering for listed senders and IP addresses, so a forged or compromised sender gets straight through.
  • DE.CM-09 Computing monitored. Allow lists skip spam filtering for listed senders and IP addresses, so a forged or compromised sender gets straight through.
  • 5.4.1 Protection against phishing. Allow lists skip spam filtering for listed senders and IP addresses, so a forged or compromised sender gets straight through.
  • 500.14(a)(2) Protection against malicious code. Allow lists skip spam filtering for listed senders and IP addresses, so a forged or compromised sender gets straight through.
  • 248.30(a)(2)(ii) Protection against anticipated threats. Allow lists skip spam filtering for listed senders and IP addresses, so a forged or compromised sender gets straight through.
  • CC6.8 Preventing malicious software. Allow lists skip spam filtering for listed senders and IP addresses, so a forged or compromised sender gets straight through.
  • CI-07 Email security. Allow lists skip spam filtering for listed senders and IP addresses, so a forged or compromised sender gets straight through.

Email from outside the organization is tagged (EXO-015)

Low Email security

Fix: Run Set-ExternalInOutlook -Enabled $true, so Outlook marks messages from outside senders.

Take the helm: step-by-step
  1. Connect to Exchange Online PowerShell: Connect-ExchangeOnline.
  2. Run Set-ExternalInOutlook -Enabled $true.
  3. Outlook shows an External tag on messages from outside senders. It can take a couple of days to appear everywhere.

Maps to:

  • 164.308(a)(5)(ii)(B) Protection from malicious software. An External tag on outside email helps people spot phishing that pretends to come from a coworker.
  • 314.4(c) Safeguards that control identified risks. An External tag on outside email helps people spot phishing that pretends to come from a coworker.
  • DE.CM-09 Computing monitored. An External tag on outside email helps people spot phishing that pretends to come from a coworker.
  • 5.4.1 Protection against phishing. An External tag on outside email helps people spot phishing that pretends to come from a coworker.
  • CC6.8 Preventing malicious software. An External tag on outside email helps people spot phishing that pretends to come from a coworker.
  • CI-07 Email security. An External tag on outside email helps people spot phishing that pretends to come from a coworker.

Admins are notified about outbound spam (EXO-019)

Low Email security

Fix: In the default outbound spam policy, notify your admins (or a shared security mailbox) when a user is blocked for sending spam, which usually means the account is compromised.

Take the helm: step-by-step
  1. Microsoft Defender portal → Email & collaboration → Policies & rules → Threat policies → Anti-spam → Anti-spam outbound policy (Default).
  2. Under Notifications, select 'Send a copy of suspicious outbound messages' and 'Notify these users and groups if a sender is blocked', and add an admin or security mailbox.

Maps to:

  • 314.4(d)(1) Test and monitor safeguards. A user blocked for sending spam usually means a hacked mailbox. Notifying admins means someone acts on it the same day.
  • DE.CM-09 Computing monitored. A user blocked for sending spam usually means a hacked mailbox. Notifying admins means someone acts on it the same day.
  • 248.30(a)(2)(ii) Protection against anticipated threats. A user blocked for sending spam usually means a hacked mailbox. Notifying admins means someone acts on it the same day.
  • CC7.2 Monitoring for attacks. A user blocked for sending spam usually means a hacked mailbox. Notifying admins means someone acts on it the same day.
  • CI-07 Email security. A user blocked for sending spam usually means a hacked mailbox. Notifying admins means someone acts on it the same day.

Zero-hour auto purge removes threats after delivery (EXO-020)

Low Email security

Fix: Keep zero-hour auto purge on for phishing, spam, and malware in the anti-spam and anti-malware policies, so messages found to be malicious later are pulled from inboxes.

Take the helm: step-by-step
  1. Microsoft Defender portal → Email & collaboration → Policies & rules → Threat policies → Anti-spam → Anti-spam inbound policy (Default) → Actions: turn on zero-hour auto purge for spam and phishing.
  2. Anti-malware → Default policy → Protection settings: turn on zero-hour auto purge for malware.

Maps to:

  • SI.L2-3.14.2 Provide protection from malicious code. Zero-hour auto purge pulls messages out of inboxes when they're identified as phishing or malware after delivery, before more people open them.
  • SI.L1-B.1.XIII Protect against malicious code. Zero-hour auto purge pulls messages out of inboxes when they're identified as phishing or malware after delivery, before more people open them.
  • 164.308(a)(5)(ii)(B) Protection from malicious software. Zero-hour auto purge pulls messages out of inboxes when they're identified as phishing or malware after delivery, before more people open them.
  • 314.4(c) Safeguards that control identified risks. Zero-hour auto purge pulls messages out of inboxes when they're identified as phishing or malware after delivery, before more people open them.
  • DE.CM-09 Computing monitored. Zero-hour auto purge pulls messages out of inboxes when they're identified as phishing or malware after delivery, before more people open them.
  • 5.4.1 Protection against phishing. Zero-hour auto purge pulls messages out of inboxes when they're identified as phishing or malware after delivery, before more people open them.
  • 500.14(a)(2) Protection against malicious code. Zero-hour auto purge pulls messages out of inboxes when they're identified as phishing or malware after delivery, before more people open them.
  • 248.30(a)(2)(ii) Protection against anticipated threats. Zero-hour auto purge pulls messages out of inboxes when they're identified as phishing or malware after delivery, before more people open them.
  • CC6.8 Preventing malicious software. Zero-hour auto purge pulls messages out of inboxes when they're identified as phishing or malware after delivery, before more people open them.
  • CI-07 Email security. Zero-hour auto purge pulls messages out of inboxes when they're identified as phishing or malware after delivery, before more people open them.

Devices

Every device platform in use has a compliance policy (INT-001)

Medium Devices

Fix: In Intune, create and assign a compliance policy for each platform your managed devices run (Windows, macOS, iOS, Android).

Take the helm: step-by-step
  1. Needs Microsoft Intune Plan 1 (included in Microsoft 365 Business Premium and E3/E5).
  2. Intune admin center → Devices → Compliance → Create policy, once for each platform listed in the evidence.
  3. Set at least: require a password, require encryption, and a minimum OS version. Assign the policy to All devices or All users.

Maps to:

  • AC.L2-3.1.18 Control connection of mobile devices. Compliance policies set the rules a phone, tablet, or laptop must meet before it connects to company data.
  • CM.L2-3.4.2 Establish and enforce security configuration settings. A compliance policy defines what a healthy device looks like, such as a passcode, encryption, and a current OS. Without one, a platform's devices are never checked.
  • 4.1 Establish and Maintain a Secure Configuration Process. A compliance policy defines what a healthy device looks like, such as a passcode, encryption, and a current OS. Without one, a platform's devices are never checked.
  • 314.4(c)(2) Know your systems and devices. A compliance policy defines what a healthy device looks like, such as a passcode, encryption, and a current OS. Without one, a platform's devices are never checked.
  • PR.PS-01 Configuration management. A compliance policy defines what a healthy device looks like, such as a passcode, encryption, and a current OS. Without one, a platform's devices are never checked.
  • 500.3(c) Asset inventory, device management, and end of life. A compliance policy defines what a healthy device looks like, such as a passcode, encryption, and a current OS. Without one, a platform's devices are never checked.
  • CC7.1 Finding vulnerabilities and configuration drift. A compliance policy defines what a healthy device looks like, such as a passcode, encryption, and a current OS. Without one, a platform's devices are never checked.
  • CI-09 Hardened configuration. A compliance policy defines what a healthy device looks like, such as a passcode, encryption, and a current OS. Without one, a platform's devices are never checked.

Devices without a compliance policy count as not compliant (INT-002)

Medium Devices

Fix: In Intune (Endpoint security, Device compliance, Compliance settings), set 'Mark devices with no compliance policy assigned as' to Not compliant.

Take the helm: step-by-step
  1. Intune admin center → Devices → Compliance → Compliance settings.
  2. Set 'Mark devices with no compliance policy assigned as' = Not compliant, then Save.

Maps to:

  • AC.L2-3.1.18 Control connection of mobile devices. When devices without a policy count as not compliant, a newly enrolled or forgotten device can't slip through as healthy.
  • 4.1 Establish and Maintain a Secure Configuration Process. When devices without a policy count as not compliant, a newly enrolled or forgotten device can't slip through as healthy.
  • 314.4(c)(2) Know your systems and devices. When devices without a policy count as not compliant, a newly enrolled or forgotten device can't slip through as healthy.
  • PR.PS-01 Configuration management. When devices without a policy count as not compliant, a newly enrolled or forgotten device can't slip through as healthy.
  • CC7.1 Finding vulnerabilities and configuration drift. When devices without a policy count as not compliant, a newly enrolled or forgotten device can't slip through as healthy.

Managed devices are compliant (INT-003)

Medium Devices

Fix: Review the noncompliant devices in Intune, fix what each one is missing, and pair compliance with a Conditional Access policy so noncompliant devices can't reach company data.

Default thresholds: Target percentage = 95

Take the helm: step-by-step
  1. Intune admin center → Devices → Monitor → Noncompliant devices. Open each device in the evidence to see which setting fails.
  2. Fix the setting on the device, or have the user follow the Company Portal's instructions.
  3. Add a Conditional Access policy that requires a compliant device, so noncompliant devices can't reach email and files until they're fixed.

Maps to:

  • AC.L2-3.1.18 Control connection of mobile devices. Compliance shows which devices actually meet your rules. Fixing noncompliant devices closes the gaps they expose.
  • CM.L2-3.4.2 Establish and enforce security configuration settings. Compliance shows which devices actually meet your rules. Fixing noncompliant devices closes the gaps they expose.
  • 164.308(a)(1)(ii)(B) Risk management. Noncompliant devices are a concrete list of risks to fix, which feeds straight into risk management.
  • 4.1 Establish and Maintain a Secure Configuration Process. Compliance shows which devices actually meet your rules. Fixing noncompliant devices closes the gaps they expose.
  • 314.4(c)(2) Know your systems and devices. Compliance shows which devices actually meet your rules. Fixing noncompliant devices closes the gaps they expose.
  • PR.PS-01 Configuration management. Compliance shows which devices actually meet your rules. Fixing noncompliant devices closes the gaps they expose.
  • 500.3(c) Asset inventory, device management, and end of life. Compliance shows which devices actually meet your rules. Fixing noncompliant devices closes the gaps they expose.
  • CC7.1 Finding vulnerabilities and configuration drift. Compliance shows which devices actually meet your rules. Fixing noncompliant devices closes the gaps they expose.
  • CI-09 Hardened configuration. Compliance shows which devices actually meet your rules. Fixing noncompliant devices closes the gaps they expose.

Intune-managed devices are encrypted (INT-004)

High Devices

Fix: Turn on encryption for each unencrypted device: BitLocker on Windows, FileVault on Mac, a passcode on iPhone and iPad, and device encryption on Android. Intune can enforce each one.

Take the helm: step-by-step
  1. Windows: an Endpoint security → Disk encryption policy that turns on BitLocker.
  2. Mac: a Disk encryption policy that turns on FileVault and escrows the recovery key in Intune.
  3. iPhone and iPad: encrypted once a passcode is set; require a passcode in the compliance policy.
  4. Android: require encryption in the compliance policy (Work profile or fully managed).

Maps to:

  • AC.L2-3.1.19 Encrypt CUI on mobile devices. Encryption keeps the data on a lost or stolen laptop or phone unreadable without the passcode or key.
  • SC.L2-3.13.16 Protect the confidentiality of CUI at rest. Encryption keeps the data on a lost or stolen laptop or phone unreadable without the passcode or key.
  • 164.312(a)(2)(iv) Encryption and decryption. Encryption keeps the data on a lost or stolen laptop or phone unreadable without the passcode or key.
  • 3.6 Encrypt Data on End-User Devices. Encryption keeps the data on a lost or stolen laptop or phone unreadable without the passcode or key.
  • 314.4(c)(3) Encryption. Encryption keeps the data on a lost or stolen laptop or phone unreadable without the passcode or key.
  • PR.DS-01 Data at rest protected. Encryption keeps the data on a lost or stolen laptop or phone unreadable without the passcode or key.
  • 500.15(a) Encryption. Encryption keeps the data on a lost or stolen laptop or phone unreadable without the passcode or key.
  • 248.30(a)(2)(i) Security and confidentiality of customer information. Encryption keeps the data on a lost or stolen laptop or phone unreadable without the passcode or key.
  • CC6.1 Logical access security. Encryption keeps the data on a lost or stolen laptop or phone unreadable without the passcode or key.
  • CI-10 Encrypted laptops. Encryption keeps the data on a lost or stolen laptop or phone unreadable without the passcode or key.

Compliance policies require a passcode and encryption (INT-005)

Medium Devices

Fix: In each platform's compliance policy, require a password or passcode and require encryption of data storage.

Take the helm: step-by-step
  1. Intune admin center → Devices → Compliance → open each policy named in the evidence → Properties → Compliance settings.
  2. Under System security (or Password and Device security), require a password and require encryption of data storage. Save.

Maps to:

  • AC.L2-3.1.18 Control connection of mobile devices. Requiring a passcode and encryption in the compliance policy means every enrolled device is held to them, not just the ones someone set up carefully.
  • AC.L2-3.1.19 Encrypt CUI on mobile devices. Requiring a passcode and encryption in the compliance policy means every enrolled device is held to them, not just the ones someone set up carefully.
  • 164.312(a)(2)(iv) Encryption and decryption. Requiring a passcode and encryption in the compliance policy means every enrolled device is held to them, not just the ones someone set up carefully.
  • 3.6 Encrypt Data on End-User Devices. Requiring a passcode and encryption in the compliance policy means every enrolled device is held to them, not just the ones someone set up carefully.
  • 314.4(c)(3) Encryption. Requiring a passcode and encryption in the compliance policy means every enrolled device is held to them, not just the ones someone set up carefully.
  • PR.DS-01 Data at rest protected. Requiring a passcode and encryption in the compliance policy means every enrolled device is held to them, not just the ones someone set up carefully.
  • 500.15(a) Encryption. Requiring a passcode and encryption in the compliance policy means every enrolled device is held to them, not just the ones someone set up carefully.
  • CC6.1 Logical access security. Requiring a passcode and encryption in the compliance policy means every enrolled device is held to them, not just the ones someone set up carefully.
  • CI-10 Encrypted laptops. Requiring a passcode and encryption in the compliance policy means every enrolled device is held to them, not just the ones someone set up carefully.

App protection policies protect company data on phones (INT-006)

Medium Devices

Fix: In Intune (Apps, App protection policies), create and assign policies for iOS/iPadOS and Android that require a PIN and block copying company data into personal apps.

Take the helm: step-by-step
  1. Intune admin center → Apps → App protection policies → Create policy → iOS/iPadOS, then again for Android.
  2. Target all Microsoft apps. Data protection: block Save copies of org data and restrict cut, copy, and paste to managed apps. Access requirements: require a PIN.
  3. Assign both policies to all users, or the groups that use phones for work.

Maps to:

  • AC.L2-3.1.18 Control connection of mobile devices. App protection keeps company data inside managed apps on phones, even personal ones, so it can't be copied to personal apps and can be wiped without touching personal data.
  • AC.L2-3.1.3 Control the flow of CUI. App protection keeps company data inside managed apps on phones, even personal ones, so it can't be copied to personal apps and can be wiped without touching personal data.
  • 164.312(a)(1) Access control. App protection keeps company data inside managed apps on phones, even personal ones, so it can't be copied to personal apps and can be wiped without touching personal data.
  • 314.4(c)(1) Access controls. App protection keeps company data inside managed apps on phones, even personal ones, so it can't be copied to personal apps and can be wiped without touching personal data.
  • PR.DS-01 Data at rest protected. App protection keeps company data inside managed apps on phones, even personal ones, so it can't be copied to personal apps and can be wiped without touching personal data.
  • 248.30(a)(2)(iii) Protection against unauthorized access. App protection keeps company data inside managed apps on phones, even personal ones, so it can't be copied to personal apps and can be wiped without touching personal data.
  • CC6.7 Controlling how data leaves. App protection keeps company data inside managed apps on phones, even personal ones, so it can't be copied to personal apps and can be wiped without touching personal data.

Managed devices check in regularly (INT-007)

Low Devices

Fix: Find devices that haven't synced with Intune in 30 days: reconnect them, or retire and delete devices that are gone.

Default thresholds: staleDays = 30

Take the helm: step-by-step
  1. Intune admin center → Devices → All devices, sort by Last check-in, and open the devices named in the evidence.
  2. Ask the user to open the Company Portal and sync. If the device is gone, select Retire, then Delete.
  3. Optionally add device clean-up rules (Devices → Device clean-up rules) to remove devices inactive for 90 days.

Maps to:

  • CM.L2-3.4.1 Maintain baseline configurations and inventories. Devices that stop checking in drop out of your view and stop getting policy. Finding them keeps the inventory accurate.
  • 1.1 Establish and Maintain Detailed Enterprise Asset Inventory. Devices that stop checking in drop out of your view and stop getting policy. Finding them keeps the inventory accurate.
  • 314.4(c)(2) Know your systems and devices. Devices that stop checking in drop out of your view and stop getting policy. Finding them keeps the inventory accurate.
  • ID.AM-01 Hardware inventory. Devices that stop checking in drop out of your view and stop getting policy. Finding them keeps the inventory accurate.
  • 500.13(a) Asset inventory. Devices that stop checking in drop out of your view and stop getting policy. Finding them keeps the inventory accurate.
  • 12.5.1 Inventory of system components. Devices that stop checking in drop out of your view and stop getting policy. Finding them keeps the inventory accurate.
  • CC6.1 Logical access security. Devices that stop checking in drop out of your view and stop getting policy. Finding them keeps the inventory accurate.