Turn findings into progress
Your action plan
Recommended fixes, grouped by effort and ordered with their prerequisites. Review each change before applying it.
Recommended actions
77One for each result that needs attentionQuick wins
3Start with a focused work sessionEstimated effort
About 27 hrFor 53 fixes; 24 not estimated. Allow time for testing and coordination.Current posture
46/100Partial assessment · 132 measured · 23 not verifiedReview → make the change in Microsoft → verify. Tenant Strike reads your settings. It does not apply these changes. Projected score improvements assume the recommended fixes work; a fresh assessment is needed to confirm the result. Planned fixes stay open until evidence shows otherwise.
Quick wins: 3 fixes
Start here: the fixes that do the most good for the least work. A fix with no time estimate is counted as a short one.
- Effort
- About 1 hr
- 1 not estimated
- Projected
- 49
- Lift
- +3
- 01SeriousIdentity
Require MFA for all users
What this means
No supported baseline for mandatory MFA for modern user sign-ins was established. 2 policies reviewed; none is an enforcing candidate for the required users or roles, all applications and client types with mandatory grants. This is a baseline gap, not proof that every sign-in is unprotected.
What to do
Review policy assignments, grant alternatives and excluded accounts with your IT administrator. Preserve emergency access; verify its separate protection instead of removing exclusions just to pass a scan.
- Who does it
- Your IT person or provider
- How long
- About 30 minutes
2 attack paths weakenedTechnical detailssetting, steps
The setting
Conditional Access policies
Entra admin center → Protection → Conditional Access → Policies
Now: Supported baseline not established→Should be: Effective scope verified, including protected emergency access
Open in the admin portal ↗Learn more ↗Steps
- Review policies and their user, role, application, client, location and device scopes.
- Check whether MFA is mandatory or an alternative to another grant control.
- Use Conditional Access What If and sign-in logs for representative users, admins and excluded accounts.
- Test changes in report-only mode and verify emergency access before enforcing them.
- 02SeriousIdentity
Enable Security Defaults or Conditional Access
What this means
Security Defaults is disabled. 2 policies reviewed. No universal MFA baseline candidate was established. A universal legacy-authentication block was not established.
What to do
Ask your IT administrator to review the existing baseline and its exceptions. Use Security Defaults or tested Conditional Access policies appropriate to the tenant.
- Who does it
- Your IT person or provider
- How long
- Not estimated
1 attack path weakenedTechnical detailssetting
Learn more ↗The setting
Identity protection baseline
Entra admin center → Protection → Conditional Access
Open in the admin portal ↗ - 03SeriousApps & Consent
Restrict user app consent
What this means
User consent is fully open — any user can grant any OAuth app any permission including Mail.ReadWrite, Files.ReadWrite.All, and Directory.ReadWrite.All. This is the worst possible consent posture and the basis for illicit-consent attacks.
What to do
Change to 'Allow user consent for apps from verified publishers, for selected permissions' AND enable the admin consent workflow.
- Who does it
- Your IT person or provider
- How long
- About 30 minutes
- Before you change it
- Users who currently consent to apps will need to request admin consent instead. Inform users and ensure a reviewer is monitoring requests.
1 attack path weakenedTechnical detailssetting, steps
The setting
User consent settings for enterprise applications
Entra admin center → Identity → Applications → Enterprise applications → Consent and permissions → User consent settings
Now: Allow user consent for apps (legacy — all permissions)→Should be: Allow user consent for apps from verified publishers, for selected permissions
Open in the admin portal ↗Learn more ↗Steps
- entra.microsoft.com → Identity → Applications → Enterprise applications → Consent and permissions
- User consent settings → 'Allow user consent for apps from verified publishers, for selected permissions'
- Admin consent requests → Set 'Users can request admin consent to apps they are unable to consent to' → Yes
- Configure reviewers for incoming consent requests
This week: 73 fixes
Fit these around other work: more short fixes, and fixes that take up to about two hours.
- Effort
- About 23 hr
- 23 not estimated
- Projected
- 99
- Lift
- +53
- 01SeriousAzure
Block anonymous container access
Subscription: Contoso Production
What this means
1 blob container in subscription "Contoso Production" has Blob or Container anonymous-access settings. Effective public access depends on the account setting and network restrictions and was not demonstrated by this check.
What to do
Set publicAccess to None on each container. If anonymous access is genuinely required, document it as an exception with rationale.
- Who does it
- Whoever runs your Azure
- How long
- Not estimated
1 attack path weakenedTechnical detailssetting, steps
The setting
Container public access level
Storage account → Containers → [container] → Change access level
Now: Container→Should be: Private (None)
Open in the admin portal ↗Steps
- portal.azure.com → Storage account → Containers → [container]
- Change access level → Private (no anonymous access) → OK
Learn more ↗PowerShell
az storage container set-permission --account-name <storage> --name <container> --public-access off
- 02SeriousAzure
Block SSH from the internet
Subscription: Contoso Production
What this means
1 NSG rule in subscription "Contoso Production" allows SSH (TCP/22) from the internet.
What to do
Restrict the source to specific IP ranges, or replace direct SSH access with Azure Bastion / VPN.
- Who does it
- Whoever runs your Azure
- How long
- Not estimated
1 attack path weakenedTechnical detailssetting, steps
The setting
Inbound SSH (TCP/22) source
Network security group → Inbound security rules → [rule]
Now: Source: * / Internet / 0.0.0.0/0→Should be: Source: specific IP range, VirtualNetwork, or use Azure Bastion
Open in the admin portal ↗Steps
- Network security group → Inbound security rules → [rule]
- Source: change from '*'/'Internet' to specific IP range or 'VirtualNetwork'
- Save. For admin access at scale, deploy Azure Bastion per VNet.
Learn more ↗PowerShell
az network nsg rule update --nsg-name <nsg> -g <rg> --name <rule> --source-address-prefixes <allowed-ip-range>
- 03SeriousAzure
Narrow SQL firewall rules
Subscription: Contoso Production
What this means
1 broad firewall rule on SQL Servers in subscription "Contoso Production".
What to do
Remove the all-IPs / all-Azure-services rule. Use VNet rules (private endpoints) or narrow IP allowlists scoped to the specific clients that need access.
- Who does it
- Whoever runs your Azure
- How long
- Not estimated
1 attack path weakenedTechnical detailssetting, steps
The setting
SQL Server firewall rules
SQL server → Security → Networking
Now: 1 broad rule (e.g. allow-all: All IPv4 addresses (0.0.0.0 → 255.255.255.255))→Should be: No 0.0.0.0/0 or AllowAllWindowsAzureIps rules; use private endpoints or narrow IP allowlists
Open in the admin portal ↗Steps
- SQL Server → Security → Networking
- Public network access: Selected networks (or Disabled if using private endpoints only)
- Remove rules with 0.0.0.0/0 → 255.255.255.255 or the 'AllowAllWindowsAzureIps' rule.
- Add specific IP ranges for the clients that legitimately need access.
Learn more ↗PowerShell
az sql server firewall-rule delete --server <server> -g <rg> --name <rule>
- 04SeriousExternal Attack Surface
Close exposed database ports
What this means
1 database endpoint(s) reachable from the public internet. This is the textbook 'leaked MongoDB' / 'open Elasticsearch' scenario — treat every one as a potential data-breach incident waiting to be discovered. Port and service data as observed on 2026-11-09 (Shodan index, not a live probe).
What to do
Block all database ports at the network edge. Databases should be accessible only from the application tier within a private network.
- Who does it
- Whoever runs your website and domain
- How long
- About 1 hour
- Before you change it
- Anything currently connecting to the database from outside your network will fail. Inventory legitimate connections (analytics tools, BI dashboards) first; route them through a VPN or bastion.
1 attack path weakenedTechnical detailssetting, steps
The setting
Public-facing database services (1 endpoint(s))
Your firewall / cloud security group / hosting provider control panel
Now: 1 database endpoint(s) (Mongo/Elasticsearch/Redis/SQL/etc.) reachable from the public internet→Should be: 0 database ports reachable from the public internet; databases reachable only from the application tier via private networking
Open in the admin portal ↗Steps
- Identify the cloud account / data center hosting each database
- Apply Network Security Group / security group / firewall rule blocking the port from the internet
- Re-test with `curl` or `nmap` from an external IP to verify
- If the application that uses the database is also cloud-hosted, use private networking (VNet/VPC peering) instead of public IPs
- If on-prem: move the DB behind your perimeter firewall, not in the DMZ
- 05SeriousExternal Attack Surface
Close exposed management ports
What this means
1 remote-management endpoint(s) reachable from the public internet. Each is a direct attack surface — credential spray, exploit kits, and ransomware affiliates routinely scan the entire IPv4 space for these. Port and service data as observed on 2026-11-09 (Shodan index, not a live probe).
What to do
Put management interfaces behind a VPN, Tailscale/ZeroTier-style mesh, or an identity-aware proxy. If business need requires public access (rare), at minimum: source-IP allowlist at the network edge AND enforce MFA at the application layer.
- Who does it
- Whoever runs your website and domain
- How long
- About 2 hours
- Before you change it
- Removing public RDP/SSH means admins working from home need a different access path (VPN, bastion, ZTNA agent). Inventory current admin usage patterns before the cutover.
1 attack path weakenedTechnical detailssetting, steps
The setting
Public-facing remote-management services (1 endpoint(s))
Your firewall / cloud security group / hosting provider control panel
Now: 1 management endpoint(s) reachable from the public internet→Should be: 0 management endpoints (RDP/SMB/Telnet/WinRM/VNC), control panels (cPanel/WHM/Control Web Panel/Webmin) or FTP reachable from the public internet; access only via VPN, bastion, or identity-aware proxy
Open in the admin portal ↗Learn more ↗Steps
- Identify the device/server hosting each exposed port (the IP + hostname above)
- At the firewall: restrict the port to known admin source IPs (VPN, jumphost) or block entirely
- For RDP specifically: deploy Azure Bastion / RDS Gateway / a SaaS jumphost instead
- For SMB: there is essentially no business reason to expose 445 — block it
- For VPN gateways: prefer cloud-managed (Entra ID conditional access in front) over device-managed
- For a hosting control panel (cPanel/WHM/Control Web Panel/Webmin): ask the host to restrict the panel to your office/VPN addresses, and never sign in over the plain-HTTP port
- For FTP: switch to SFTP or FTPS and turn off plain FTP — most hosts support this already
- 06SeriousDevices
Require BitLocker on Windows
What this means
2 Windows devices report BitLocker disabled. Each unencrypted device is a data-breach incident waiting for a theft or loss.
What to do
Make BitLocker mandatory via a compliance policy setting. Devices that fail will be marked non-compliant and (with a CA policy) blocked from corp resources until encrypted.
- Who does it
- Your IT person or provider
- How long
- About 1.5 hours
- Before you change it
- Devices currently unencrypted will be marked non-compliant. Users will need to enable BitLocker (15-60 min per device depending on disk size). Push the policy with a grace period so users have time.
3 attack paths weakenedTechnical detailssetting, steps
The setting
BitLocker / disk encryption policy
Intune admin center → Endpoint security → Disk encryption
Now: 2 Windows devices with BitLocker disabled, 0 with unknown status (of 4 total)→Should be: BitLocker required by compliance policy and enabled on all Windows devices
Open in the admin portal ↗Learn more ↗Steps
- intune.microsoft.com → Devices → Compliance policies
- Open your Windows compliance policy (or create one)
- System Security → Encryption → 'Require encryption of data storage' → Require
- Save
- Then: Devices → Configuration profiles → Create profile → Windows → Endpoint protection → BitLocker base settings → Enable
- 07SeriousIdentity
Require MFA for admins
What this means
No supported baseline for mandatory MFA for the 14 reference administrator roles (every role Microsoft marks privileged is one of the 14 reference roles) was established. 2 policies reviewed; none is an enforcing candidate for the required users or roles, all applications and client types with mandatory grants. Policy "Require MFA for admins" would meet it but leaves out 1 reference administrator role (Conditional Access Administrator). This is a baseline gap, not proof that every sign-in is unprotected.
What to do
Review policy assignments, grant alternatives and excluded accounts with your IT administrator. Preserve emergency access; verify its separate protection instead of removing exclusions just to pass a scan.
- Who does it
- Your IT person or provider
- How long
- About 30 minutes
Technical detailssetting, steps
The setting
Conditional Access policies
Entra admin center → Protection → Conditional Access → Policies
Now: Supported baseline not established→Should be: Effective scope verified, including protected emergency access
Open in the admin portal ↗Learn more ↗Steps
- Review policies and their user, role, application, client, location and device scopes.
- Check whether MFA is mandatory or an alternative to another grant control.
- Use Conditional Access What If and sign-in logs for representative users, admins and excluded accounts.
- Test changes in report-only mode and verify emergency access before enforcing them.
- 08SeriousExternal Attack Surface
Certificates past their expiry date
What this means
1 certificate(s) on record have passed their expiry date. The record may be older than this scan, so a certificate may have been renewed since. If it has not, browsers and apps refuse to connect to that service. Port and service data as observed on 2026-11-09 (Shodan index, not a live probe).
What to do
Check each service. If its certificate has not been renewed, renew it now and fix whatever should have renewed it.
- Who does it
- Whoever runs your website and domain
- How long
- About 30 minutes
Technical detailssetting
The setting
TLS certificate(s) past their expiry date on public services (1)
The hosting environment / load balancer / CDN serving each affected service (ACM, Let's Encrypt automation, IIS, nginx, etc.)
Now: 1 certificate(s) past their expiry date on record→Should be: 0 expired certificates; auto-renewal job verified green
Open in the admin portal ↗ - 09SeriousAzure
Block RDP from the internet
Subscription: Contoso Production
What this means
1 NSG rule in subscription "Contoso Production" allows RDP (TCP/3389) from the internet.
What to do
Restrict the source to specific IP ranges, or replace with Azure Bastion. Ransomware crews commonly get in this way, so fix this first if you see it.
- Who does it
- Whoever runs your Azure
- How long
- Not estimated
Technical detailssetting, steps
The setting
Inbound RDP (TCP/3389) source
Network security group → Inbound security rules → [rule]
Now: Source: * / Internet / 0.0.0.0/0→Should be: Source: specific IP range, or remove rule and use Azure Bastion
Open in the admin portal ↗Steps
- Network security group → Inbound security rules → [rule]
- Source: change to specific IP range or remove the rule
- Deploy Azure Bastion for admin access (no RDP exposed publicly).
Learn more ↗PowerShell
az network nsg rule update --nsg-name <nsg> -g <rg> --name <rule> --source-address-prefixes <allowed-ip>
- 10SeriousDevices
Fix noncompliant devices
What this means
2 of 5 devices (40%) are actively non-compliant. Each represents a control you've defined that's currently being violated.
What to do
Review each non-compliant device: address the specific failing setting (disk encryption, OS version, antivirus, etc.) or remove the device from Intune if it's no longer in use.
- Who does it
- Your IT person or provider
- How long
- About 1 hour
1 other fix unblockedTechnical detailssetting, steps
The setting
Intune device compliance status
Intune admin center → Devices → All devices (filter by Compliance: Not compliant)
Now: 2 non-compliant, 0 in grace period, 0 not evaluated (of 5 enrolled)→Should be: 0 devices in noncompliant / conflict / error state
Open in the admin portal ↗Steps
- intune.microsoft.com → Reports → Device compliance → Noncompliant devices
- Drill into each device → see specific policy items failing
- Common fixes: re-encrypt disk, update OS, enable Defender, restart agent
Learn more ↗PowerShell
# Trigger a sync from PowerShell on a Windows device: Get-ScheduledTask -TaskPath "\Microsoft\Windows\EnterpriseMgmt\*" | Where-Object {$_.TaskName -eq 'PushLaunch'} | Start-ScheduledTask - 11Should fixDevices
Set compliance policies per platform
What this means
1 OS platform with enrolled devices has no assigned compliance policy: iOS. Devices on these platforms cannot be evaluated for compliance, so CA policies requiring 'compliant device' will treat them inconsistently.
What to do
Create a compliance policy for each missing platform. Start with the Intune compliance baseline as a template.
- Who does it
- Your IT person or provider
- How long
- About 30 minutes
- Before you change it
- Devices already enrolled that fail the new policy will be marked non-compliant. Use a grace period (Actions for noncompliance) so users have 7+ days to remediate before access is blocked.
1 other fix unblockedTechnical detailssetting, steps
The setting
Intune compliance policies
Intune admin center → Devices → Compliance policies → Policies
Now: No assigned compliance policy for: iOS→Should be: At least one compliance policy assigned for every enrolled OS platform
Open in the admin portal ↗Learn more ↗Steps
- intune.microsoft.com → Devices → Compliance policies → Create policy
- Platform: select the missing platform → Create
- Use the Microsoft baseline as a starting point
- Assign to All users / All devices
- 12SeriousDevices
Require compliant devices
What this means
No Conditional Access policy requires a compliant device for any application. Your compliance evaluation is purely descriptive — it tells you who's non-compliant but doesn't prevent non-compliant (or attacker-controlled, unenrolled) devices from accessing corp data.
What to do
Create a CA policy requiring 'Compliant device' (or 'Hybrid Azure AD joined device') for access to all cloud apps, or at minimum for sensitive apps (Exchange, SharePoint, finance apps).
- Who does it
- Your IT person or provider
- How long
- About 45 minutes
- Before you change it
- Users on non-Intune-enrolled devices will be blocked from sign-in. This is the intended behavior but will affect: BYOD users without Intune enrollment, kiosks, contractors with their own devices. Plan exceptions or alternative paths (MAM-only, App Protection Policies, etc.) before enforcing.
2 attack paths weakenedBefore this step: Set compliance policies per platform, Fix noncompliant devicesTechnical detailssetting, steps
The setting
Conditional Access — require compliant device grant
Entra admin center → Protection → Conditional Access → Policies
Now: 1 CA policy enabled, none requiring compliant or hybrid-joined device→Should be: At least one enabled CA policy granting access only when the device is marked compliant or hybrid-joined
Open in the admin portal ↗Learn more ↗Steps
- entra.microsoft.com → Protection → Conditional Access → New policy
- Name: 'Require compliant device — all apps'
- Users: All users; Exclude: break-glass account, service accounts that use modern auth from non-Intune systems
- Cloud apps: All cloud apps (or scope to Exchange/SharePoint first)
- Grant: Grant access → Require device to be marked as compliant (and/or Require Hybrid Azure AD joined device) → Require one of the selected controls
- Enable in Report-only for 2 weeks first, review impact, then switch On
- 13SeriousEmail
Enforce DMARC
What this means
contoso.com: DMARC is published but set to p=none for contoso.com — monitoring mode only. Spoofed email is not blocked. Move to p=quarantine after reviewing reports. Also checked: contoso.onmicrosoft.com (warn).
What to do
Move from p=none → p=quarantine after reviewing aggregate reports.
- Who does it
- Your IT person or provider
- How long
- About 10 minutes
1 attack path weakenedTechnical details
- 14SeriousAzure
Register MFA for everyone who controls Azure
What this means
1 account that can control your Azure resources has no MFA method registered, so a stolen or guessed password is enough to sign in to it. That access covers a whole subscription. 1 guest also holds these roles. A guest's MFA is set up in their own organization, so it is not judged here; "Review guest role assignments" lists guests with a role of their own.
What to do
Have the person behind each account listed sign in and set up an MFA method, such as Microsoft Authenticator or a passkey. Remove the role from any account that no longer needs it.
- Who does it
- Whoever runs your Azure
- How long
- About 10 minutes
1 attack path weakenedTechnical detailssetting, steps
The setting
MFA methods of accounts with Azure roles
Microsoft Entra admin center → Users → [account] → Authentication methods
Now: 1 account with an Azure role has no MFA method→Should be: Every account with an Azure role has an MFA method
Open in the admin portal ↗Learn more ↗Steps
- Ask the person behind each account listed to sign in at https://aka.ms/mysecurityinfo and add Microsoft Authenticator or a passkey.
- To confirm: Microsoft Entra admin center → Users → [account] → Authentication methods.
- If an account no longer needs the access: Azure portal → Subscriptions → [subscription] → Access control (IAM) → Role assignments, and remove it.
- If a script or a device signs in with the account, not a person: move that job to a service principal or managed identity, which Microsoft recommends for automation.
- 15SeriousEmail
No DKIM signing key published
What this means
No DKIM signing key is published for contoso.com under Microsoft 365's two names, selector1 and selector2. Mail from contoso.com is not signed in your own name. Microsoft 365 still signs the mail it sends, but under your onmicrosoft.com name, which does not count for DMARC. So DMARC for contoso.com rests on SPF alone, and SPF breaks when mail is forwarded.
What to do
Enable DKIM signing for the domain in the Defender portal — this publishes the two required CNAMEs (you publish them at your DNS provider) and turns on signing.
- Who does it
- Your IT person or provider
- How long
- About 20 minutes
1 attack path weakenedTechnical detailssetting, steps
The setting
DKIM signing for contoso.com
Microsoft Defender portal → Email & collaboration → Policies & rules → Threat policies → Email authentication settings → DKIM
Now: No signing key published under selector1 or selector2→Should be: Enabled — both selector1 and selector2 CNAMEs published, signing toggled ON
Open in the admin portal ↗Learn more ↗Steps
- security.microsoft.com → Email & collaboration → Policies & rules → Threat policies → Email authentication settings → DKIM
- Select contoso.com
- Microsoft shows you the two CNAME records — publish them at your DNS provider
- Return to portal and toggle 'Sign messages for this domain with DKIM signatures' to ON
- 16SeriousAzure
Back up virtual machines
Subscription: Contoso Production
What this means
In subscription "Contoso Production": 1 of 1 virtual machine has no Azure Backup protection at all. No Recovery Services vault or Backup vault exists anywhere in the scanned subscriptions, so there is no Azure-native restore point for any of these machines. If these machines are protected by a third-party backup product such as Veeam, Datto or Rubrik, that protection leaves no trace in Azure Resource Manager and is invisible to this scan — in that case this finding can be dismissed.
What to do
Create (or reuse) a Recovery Services vault in the same region and enable backup on each machine with a daily policy. For machines that already have a policy, repair the stopped or failing job rather than re-enrolling.
- Who does it
- Whoever runs your Azure
- How long
- About 20 minutes
- Before you change it
- Backup adds storage cost proportional to the protected data and retention period, and the first full backup places load on the VM's disks. Nothing about the running workload changes.
1 attack path weakenedTechnical detailssetting, steps
The setting
Backup
Virtual machine → Operations → Backup
Now: Not configured→Should be: Enabled, with a healthy daily recovery point
Open in the admin portal ↗Learn more ↗Steps
- portal.azure.com → Business Continuity Center (or Backup center) → Backup
- Datasource type: Azure Virtual machines → select or create a Recovery Services vault
- Choose a backup policy (the default daily policy retains 30 days) → Enable backup
- For machines that already show a policy: Backup → Backup jobs, and resolve the failing or stopped job
- 17SeriousIdentity
Block legacy authentication
What this means
No supported baseline for a legacy-authentication block was established. 2 policies reviewed; none is an enforcing candidate for the required users or roles, all applications and client types with mandatory grants. This is a baseline gap, not proof that every sign-in is unprotected.
What to do
Review policy assignments, grant alternatives and excluded accounts with your IT administrator. Preserve emergency access; verify its separate protection instead of removing exclusions just to pass a scan.
- Who does it
- Your IT person or provider
- How long
- About 30 minutes
1 attack path weakenedTechnical detailssetting, steps
The setting
Conditional Access policies
Entra admin center → Protection → Conditional Access → Policies
Now: Supported baseline not established→Should be: Effective scope verified, including protected emergency access
Open in the admin portal ↗Learn more ↗Steps
- Review policies and their user, role, application, client, location and device scopes.
- Check whether MFA is mandatory or an alternative to another grant control.
- Use Conditional Access What If and sign-in logs for representative users, admins and excluded accounts.
- Test changes in report-only mode and verify emergency access before enforcing them.
- 18SeriousIdentity
Block device code flow
What this means
Nothing in this tenant blocks device code flow. An attacker starts the flow, sends the victim a genuine microsoft.com sign-in page and a genuine 8-character code, and the victim — who sees no fake domain and no suspicious attachment — completes MFA for them. The access token lands with the attacker, already MFA-satisfied.
What to do
Create a Conditional Access policy blocking the device code flow for all users, excluding only shared-device accounts that genuinely need it.
- Who does it
- Your IT person or provider
- How long
- About 30 minutes
- Before you change it
- Hardware that has no browser and signs in by device code — Teams Rooms panels, some conference-room systems, printers and IoT devices — will stop authenticating. Filter sign-in logs on the device-code flow before enforcing, and put any accounts you find into an excluded group rather than leaving the flow open tenant-wide.
1 attack path weakenedTechnical detailssetting, steps
The setting
Conditional Access — block device code flow
Entra admin center → Protection → Conditional Access → Policies
Now: No enabled policy blocks the deviceCodeFlow transfer method→Should be: An enabled policy: All users, All resources, Authentication flows = Device code flow, Grant: Block access
Open in the admin portal ↗Learn more ↗Steps
- entra.microsoft.com → Protection → Conditional Access → Policies → New policy
- Name: 'Block device code flow'
- Assignments → Users: All users (Exclude: break-glass, and any shared-device account group)
- Assignments → Target resources: All resources
- Conditions → Authentication flows → Configure: Yes → select 'Device code flow'
- Access controls → Grant: Block access
- Enable in Report-only briefly to confirm nothing legitimate uses it, then switch to On
- Microsoft also ships this as a managed policy named 'Block device code flow' — if it is present in your list, turning that one on is enough
- 19SeriousEmail
SPF includes a name that has no SPF record
What this means
SPF issues for contoso.com: Record ends in ~all (soft fail) — receivers typically deliver but mark as suspicious. Move to -all for stronger protection. spf.protection.outlook.com is named in the record and has no SPF record, or more than one. A receiver that gets as far as that name returns PermError, and SPF effectively fails open.
What to do
Adjust SPF record. Test changes with a tool like dmarcian SPF Surveyor before publishing.
- Who does it
- Your IT person or provider
- How long
- Not estimated
1 attack path weakenedTechnical detailssetting
Learn more ↗The setting
SPF TXT record at contoso.com
Your DNS provider (GoDaddy / Cloudflare / Route 53 / etc.) → DNS records → Edit TXT record
Now: v=spf1 include:spf.protection.outlook.com ~all→Should be: v=spf1 include:spf.protection.outlook.com -all (≤10 DNS lookups)
Open in the admin portal ↗ - 20SeriousAzure
Limit subscription Owners
Subscription: Contoso Production
What this means
7 Owners on subscription "Contoso Production" exceeds Microsoft's recommended maximum of 3. Any one of these accounts, if compromised, gives an attacker full control of the subscription. People who are only eligible for Owner through PIM are not counted.
What to do
Review each Owner and demote to a scoped role (Contributor, User Access Administrator) where possible.
- Who does it
- Whoever runs your Azure
- How long
- Not estimated
1 attack path weakenedTechnical detailssetting, steps
The setting
Owner role assignments
Azure portal → Subscriptions → Contoso Production → Access control (IAM) → Role assignments
Now: 7 Owners→Should be: At most 3 Owners
Open in the admin portal ↗Learn more ↗Steps
- portal.azure.com → Subscriptions → Contoso Production → Access control (IAM)
- Filter to role: Owner
- For each assignment, evaluate if a less-privileged role suffices (Contributor for ops, Reader for monitoring, User Access Administrator for IAM-only).
- 21SeriousAzure
Restrict storage network access
Subscription: Contoso Production
What this means
1 storage account in subscription "Contoso Production" allows access from all networks.
What to do
Set the network firewall default action to Deny, add specific IP / VNet allowlists, and remove any IP rule that covers every address.
- Who does it
- Whoever runs your Azure
- How long
- Not estimated
1 attack path weakenedTechnical detailssetting, steps
The setting
Firewall default action
Storage account → Networking → Firewalls and virtual networks
Now: Allow (all networks)→Should be: Deny (selected networks)
Open in the admin portal ↗Learn more ↗Steps
- Storage account → Networking → Firewalls and virtual networks
- Public network access → Enabled from selected virtual networks and IP addresses
- Add the IPs / VNets that legitimately need access → Save
- 22SeriousAzure
Restrict Key Vault network access
Subscription: Contoso Production
What this means
1 Key Vault in subscription "Contoso Production" accepts connections from any network.
What to do
Limit each Key Vault to the networks that need it, or disable public network access and use private endpoints.
- Who does it
- Whoever runs your Azure
- How long
- Not estimated
1 attack path weakenedTechnical detailssetting, steps
The setting
Public network access
Key Vault → Networking
Now: Allow public access from all networks→Should be: Specific virtual networks and IP addresses, or disabled with private endpoints
Open in the admin portal ↗Learn more ↗Steps
- Key Vault → Networking → Firewalls and virtual networks
- Allow public access from specific virtual networks and IP addresses, and add the ones that need the vault
- Or disable public access and configure private endpoints under Private endpoint connections.
- 23SeriousIdentity
Restrict guest invitations
What this means
ANY user (including guests themselves) can invite new guests. A single compromised account or rogue guest can create persistent attacker-controlled guest accounts. This is the loosest possible setting.
What to do
Restrict guest invitations to admins (or to a 'Guest Inviter' role assigned to specific business users) and ensure the default guest role is 'Guest user' (not 'User').
- Who does it
- Your IT person or provider
- How long
- About 10 minutes
- Before you change it
- Business users who currently invite their own guests will need to request invitations through IT or via the Guest Inviter role. Communicate the change to avoid frustration.
Technical detailssetting, steps
The setting
External collaboration settings — guest invitations
Entra admin center → Identity → External Identities → External collaboration settings
Now: Guest invite setting: Anyone in the organization can invite guest users including guests and non-members→Should be: 'Only users assigned to specific admin roles can invite guest users' and default guest role 'Guest user'
Open in the admin portal ↗Learn more ↗Steps
- entra.microsoft.com → Identity → External Identities → External collaboration settings
- Guest invite settings: select 'Only users assigned to specific admin roles can invite guest users'
- Guest user access: select 'Guest user access is restricted to properties and memberships of their own directory objects'
- 24SeriousAzure
Require HTTPS on App Services
Subscription: Contoso Production
What this means
1 App Service in subscription "Contoso Production" allows plaintext HTTP.
What to do
Toggle 'HTTPS Only' to On for each App Service.
- Who does it
- Whoever runs your Azure
- How long
- Not estimated
Technical detailssetting, steps
The setting
HTTPS Only
App Service → Settings → TLS/SSL settings → Protocol settings
Now: Off→Should be: On
Open in the admin portal ↗Learn more ↗PowerShell
az webapp update --name <name> --resource-group <rg> --https-only true
- 25Should fixEmail
Turn on anti-phishing protection
What this means
Judged: Office365 AntiPhish Default. Office365 AntiPhish Default: EnableOrganizationDomainsProtection, EnableTargetedUserProtection and EnableMailboxIntelligenceProtection are all false, so no impersonation protection covers your people or domains. 1 to tighten.
What to do
Fix: impersonation protection. Add your executives and your own domain to impersonation protection.
- Who does it
- Your IT person or provider
- How long
- About 15 minutes
2 attack paths weakenedTechnical detailssetting, steps
The setting
Anti-phishing policy
Microsoft Defender portal → Email & collaboration → Policies & rules → Threat policies → Anti-phishing
Now: impersonation protection = off→Should be: Enabled = true; EnableSpoofIntelligence = true; HonorDmarcPolicy = true; at least one impersonation protection on
Open in the admin portal ↗PowerShell
Connect-ExchangeOnline Get-AntiPhishPolicy | Format-List Name, Enabled, EnableSpoofIntelligence, HonorDmarcPolicy, EnableOrganizationDomainsProtection, EnableTargetedUserProtection, EnableMailboxIntelligenceProtection
- 26Should fixIdentity
MFA registration is not protected
What this means
Anyone who gets the password of an account that has no MFA yet can register their own phone as its second step, and keep signing in as that person. No Conditional Access policy in force protects registering sign-in methods.
What to do
Add a Conditional Access policy that requires an existing second step, or a one-time pass you issue to new staff, before anyone registers a sign-in method from outside your trusted locations.
- Who does it
- Your IT person or provider
- How long
- About 30 minutes
- Before you change it
- Someone with no MFA method yet cannot register one from outside your trusted locations without a Temporary Access Pass. Guests are left out because a Temporary Access Pass does not work for them.
2 attack paths weakenedTechnical detailssetting, steps
The setting
Conditional Access: secure security info registration
Entra admin center → Protection → Conditional Access → Policies
Now: No enforcing policy targets Register security information→Should be: An enabled policy: All users (guests and emergency access accounts excluded), User actions = Register security information, outside trusted locations, Grant: multifactor authentication
Open in the admin portal ↗Learn more ↗Steps
- Entra admin center → Protection → Conditional Access → Create new policy from templates → Securing security info registration (or New policy)
- Users: All users; exclude All guest and external users, and your emergency access accounts
- Target resources → User actions → Register security information
- Conditions → Locations: include Any location, exclude All trusted locations
- Grant → Require authentication strength → Multifactor authentication
- Turn it on in Report-only first, issue a Temporary Access Pass to anyone who has not registered yet, then switch it On
- 27Should fixDevices
Onboard devices to Defender
What this means
4 Windows devices are enrolled in Intune but no Defender for Endpoint or Defender for Business license is detected on the tenant. Endpoints run with built-in AV only — no EDR signal feeds your security team.
What to do
Add Defender for Business (bundled with M365 Business Premium) or Defender for Endpoint Plan 1/2.
- Who does it
- Your IT person or provider
- How long
- Not estimated
2 attack paths weakenedTechnical detailssetting
Learn more ↗The setting
Defender for Endpoint / Business licensing
Microsoft 365 admin center → Billing → Purchase services (then onboard in Microsoft Defender portal → Settings → Endpoints)
Now: 4 Windows devices enrolled; no Defender for Endpoint or Defender for Business licence detected in the tenant→Should be: Defender for Business or Defender for Endpoint P1/P2 licensed and provisioned
Open in the admin portal ↗ - 28Should fixApps & Consent
Enable the admin consent workflow
What this means
Admin consent workflow is DISABLED. If you've restricted user consent (recommended), users currently have no in-product way to request admin approval — they'll either go without legitimate apps or look for workarounds (personal accounts, shadow IT).
What to do
Enable the admin consent workflow and assign reviewers (e.g. your IT or security team).
- Who does it
- Your IT person or provider
- How long
- About 10 minutes
1 attack path weakenedTechnical detailssetting, steps
The setting
Admin consent workflow
Entra admin center → Identity → Applications → Enterprise applications → Consent and permissions → Admin consent settings
Now: Disabled→Should be: Enabled with reviewers assigned
Open in the admin portal ↗Learn more ↗Steps
- entra.microsoft.com → Identity → Applications → Enterprise applications → Consent and permissions → Admin consent settings
- Set 'Users can request admin consent to apps they are unable to consent to' → Yes
- Reviewers: add your IT admin group or specific reviewers
- Selected users will receive email notifications for requests: Yes
- Reminders enabled: Yes
- Consent request expires after: 30 days
- 29Should fixEmail
Detected phishing is not quarantined
What this means
Judged: Default. Mail Microsoft detects as phishing goes to people's Junk Email folder, where they can still open it, instead of quarantine. 1 to tighten.
What to do
Set the phishing action to quarantine.
- Who does it
- Your IT person or provider
- How long
- About 15 minutes
1 attack path weakenedTechnical detailssetting, steps
The setting
Anti-spam inbound policy
Microsoft Defender portal → Email & collaboration → Policies & rules → Threat policies → Anti-spam → Anti-spam inbound policy
Now: PhishSpamAction = MoveToJmf→Should be: AllowedSenderDomains and AllowedSenders empty; PhishSpamAction = Quarantine; HighConfidencePhishAction = Quarantine; SpamAction and HighConfidenceSpamAction = MoveToJmf or Quarantine
Open in the admin portal ↗Learn more ↗PowerShell
Connect-ExchangeOnline Get-HostedContentFilterPolicy | Format-List Name, AllowedSenderDomains, AllowedSenders, SpamAction, HighConfidenceSpamAction, PhishSpamAction, HighConfidencePhishAction
- 30Should fixSharing & Collaboration
Restrict OneDrive sync to domain-joined devices
What this means
OneDrive sync is permitted on ANY device. Users can install OneDrive on personal laptops, home desktops, or borrowed machines and pull a full local copy of all files they can access. That data then escapes all tenant controls — BitLocker, remote wipe, CA, audit logging.
What to do
Limit OneDrive sync to devices joined to your on-premises domains, or hold sync to compliant devices with a Conditional Access policy.
- Who does it
- Your IT person or provider
- How long
- About 15 minutes
- Before you change it
- Users on personal/BYOD devices will lose the ability to sync OneDrive locally. They can still access files via OneDrive web. For BYOD-heavy organizations, consider Intune App Protection Policies as an alternative.
1 attack path weakenedTechnical detailssetting, steps
The setting
OneDrive sync restriction to domain-joined computers
SharePoint admin center → Settings → Sync → 'Allow syncing only on computers joined to specific domains'
Now: Disabled (any device can sync)→Should be: Enabled (listed domains only)
Open in the admin portal ↗Steps
- admin.microsoft.com → SharePoint admin center → Settings → Sync
- 'Allow syncing only on computers joined to specific domains' OR
- Set 'isUnmanagedSyncAppForTenantRestricted' via Graph PATCH
Learn more ↗Microsoft Graph
PATCH /beta/admin/sharepoint/settings { "isUnmanagedSyncAppForTenantRestricted": true } - 31Should fixSharing & Collaboration
Restrict Teams external access
What this means
Judged: Global. Global: AllowFederatedUsers is true and AllowedDomains is empty, so Teams external access is open to every Teams organization. 1 to tighten.
What to do
Limit Teams external access to the partner domains you work with, and block trial organizations.
- Who does it
- Your IT person or provider
- How long
- About 15 minutes
1 attack path weakenedTechnical detailssetting, steps
The setting
Teams external access: allowed domains
Teams admin center → Users → External access → Teams and Skype for Business users in external organizations
Now: AllowFederatedUsers is true and AllowedDomains is empty, so Teams external access is open to every Teams organization→Should be: Allow only specific external domains; ExternalAccessWithTrialTenants = Blocked
Open in the admin portal ↗PowerShell
Connect-MicrosoftTeams Get-CsTenantFederationConfiguration | Format-List AllowFederatedUsers, AllowedDomains, ExternalAccessWithTrialTenants # Choose the partner domains in the Teams admin center, then: Set-CsTenantFederationConfiguration -ExternalAccessWithTrialTenants Blocked
- 32Should fixIdentity
Admins can stay signed in for weeks
What this means
Your administrators can stay signed in for weeks. If someone steals a signed-in session, it keeps working until it expires. No Conditional Access policy limits how long their sessions last.
What to do
Make admin sessions end after a working day: a Conditional Access policy for your admin roles, on all resources, that makes them sign in again every few hours (no more than 24) and never keeps their browser signed in.
- Who does it
- Your IT person or provider
- How long
- About 20 minutes
- Before you change it
- Admins are asked to sign in again more often, and a browser no longer keeps them signed in after it is closed. Nothing else changes for other users.
1 attack path weakenedTechnical detailssetting, steps
The setting
Conditional Access: sign-in frequency and persistent browser session for administrators
Entra admin center → Protection → Conditional Access → Policies
Now: No sign-in frequency or persistent browser session set→Should be: An enabled policy: admin roles (or All users), All resources, Session: Sign-in frequency 24 hours or less, Persistent browser session Never persistent
Open in the admin portal ↗Learn more ↗Steps
- Entra admin center → Protection → Conditional Access → New policy
- Users: Directory roles → the same admin roles your admin MFA policy covers (exclude your emergency access accounts)
- Target resources: All resources
- Session → Sign-in frequency → Periodic reauthentication → for example 8 hours (no more than 24)
- Session → Persistent browser session → Never persistent
- Turn it on in Report-only first, check the sign-in logs, then switch it On
- 33Should fixIdentity
User MFA registration gaps
What this means
MFA is not registered for 2 of 9 assessed non-admin member accounts (22%) in Microsoft's report. This is a registration gap, not proof that a password-only sign-in would succeed.
What to do
Review these accounts and help active users register an appropriate MFA method. Check enforcement separately and preserve tested emergency access.
- Who does it
- Your IT person or provider
- How long
- About 30 minutes
1 attack path weakenedTechnical detailssetting
Learn more ↗The setting
MFA registration
Entra admin center → Protection → Authentication methods → Registration campaign
Now: 2 of 9 assessed non-admin member accounts not registered→Should be: Active accounts have suitable registered authentication methods
Open in the admin portal ↗ - 34Should fixApps & Consent
Long-lived app credentials
What this means
1 app credential(s) have validity longer than 1 year. Industry guidance is to rotate every 12 months; long-lived secrets that leak (via committed source code, support tickets, screenshots) remain exploitable for years.
What to do
Replace with shorter-lived secrets (≤ 12 months) or, better, federated identity credentials (no secret at all).
- Who does it
- Your IT person or provider
- How long
- About 30 minutes
1 attack path weakenedTechnical detailssetting
Learn more ↗The setting
App registration credential lifetime
Entra admin center → Identity → Applications → App registrations → (each app) → Certificates & secrets
Now: 1 credential(s) with >365-day validity→Should be: ≤ 12 months validity, or federated identity credentials
Open in the admin portal ↗ - 35Should fixExternal Attack Surface
Mail-capable lookalike domains
What this means
1 domain close to contoso.com was registered in the last year and is set up for email. New lookalikes set up for email are a common start to invoice fraud. Check that you recognise each one.
What to do
Review each mail-capable lookalike. If it's not yours: report it to the registrar / hosting provider for abuse, and make sure your own DMARC is at p=reject so spoofing your EXACT domain fails (note: DMARC doesn't stop lookalike domains — user awareness and monitoring do). Consider defensively registering the closest lookalikes and common alternate TLDs.
- Who does it
- Whoever runs your website and domain
- How long
- About 30 minutes
1 attack path weakenedTechnical detailssetting
The setting
Mail-capable lookalike domains (1)
Registrar abuse reporting for each domain; your DMARC policy; defensive registration
Now: 1 lookalike domain (names withheld in this sample)→Should be: No unaffiliated mail-capable domain impersonates yours
Open in the admin portal ↗ - 36Should fixAzure
Limit service principal scope
Subscription: Contoso Production
What this means
Subscription "Contoso Production": 1 app holds Contributor over all of it. An app's secret or certificate can leak, and whoever has it gets this access.
What to do
Demote each service principal to a scoped role on only the resource group(s) it actually needs. Use Reader for monitoring-only ones. Remove User Access Administrator and Role Based Access Control Administrator unless the app's job is to manage access.
- Who does it
- Whoever runs your Azure
- How long
- Not estimated
1 attack path weakenedTechnical detailssetting, steps
The setting
Service principal role assignments (subscription scope and above)
Azure portal → Subscriptions → Contoso Production → Access control (IAM) → Role assignments → filter Type=Service principal
Now: 1 app with Contributor over the whole subscription→Should be: Service principals scoped to specific resource group(s) — none holds these roles at the subscription, a management group or the root
Open in the admin portal ↗Learn more ↗Steps
- portal.azure.com → Subscriptions → Contoso Production → Access control (IAM)
- Filter by Type: Service principal
- For each one listed: remove the assignment and re-assign at the specific resource group it needs. An assignment held at a management group or the root is removed there, not on the subscription.
- 37Should fixAzure
Enable Key Vault purge protection
Subscription: Contoso Production
What this means
1 Key Vault in subscription "Contoso Production" lacks purge protection.
What to do
Enable both soft delete (retains deleted vaults/keys for 90 days) AND purge protection (prevents force-delete during retention).
- Who does it
- Whoever runs your Azure
- How long
- Not estimated
1 attack path weakenedTechnical detailssetting, steps
The setting
Soft delete + purge protection
Key Vault → Properties
Now: Soft delete on, purge protection: off→Should be: Both enabled
Open in the admin portal ↗Steps
- Key Vault → Properties
- Soft-delete → Enabled (cannot be disabled once on — Microsoft default since 2020)
- Purge protection → Enabled (irreversible — confirm before enabling on prod vaults)
Learn more ↗PowerShell
az keyvault update --name <vault> --enable-purge-protection true
- 38Should fixAzure
Alert on critical operations
Subscription: Contoso Production
What this means
Subscription "Contoso Production" has no Activity Log alert for role assignment changes or for security and policy changes.
What to do
Create an Activity Log alert for each kind of change listed, with this subscription in its scope, and send it to an action group that emails the people who look after security.
- Who does it
- Whoever runs your Azure
- How long
- Not estimated
1 attack path weakenedTechnical detailssetting, steps
The setting
Activity Log alert rules
Azure portal → Monitor → Alerts → + Create → Alert rule
Now: No alert for role assignment changes or for security and policy changes→Should be: An alert for role assignment changes and one for security and policy changes
Open in the admin portal ↗Steps
- portal.azure.com → Monitor → Alerts → + Create → Alert rule
- Scope: this subscription
- Condition: Signal type = Activity Log → choose operation → save
- Action group: email/SMS/webhook to security on-call
Learn more ↗PowerShell
# Repeat per critical operation: az monitor activity-log alert create -n "Alert-${operation}" -g <rg> --scope /subscriptions/<sub-id> --condition category=Administrative operationName=${operation} - 39Should fixAzure
Forward the Activity Log
Subscription: Contoso Production
What this means
Subscription "Contoso Production" does not send its Activity Log anywhere, so Azure deletes it after 90 days.
What to do
Add a diagnostic setting to the subscription that sends at least the Administrative and Security categories to a Log Analytics workspace, a storage account or an event hub.
- Who does it
- Whoever runs your Azure
- How long
- Not estimated
1 attack path weakenedTechnical detailssetting, steps
The setting
Subscription diagnostic setting (Activity Log export)
Azure portal → Subscriptions → Contoso Production → Activity log → Export Activity Logs → Add diagnostic setting
Now: No diagnostic setting→Should be: At least one diagnostic setting that sends the Administrative and Security categories to a Log Analytics workspace, a storage account or an event hub
Open in the admin portal ↗Steps
- portal.azure.com → Subscriptions → Contoso Production → Activity log
- Export Activity Logs (in the top toolbar)
- Add diagnostic setting → tick every category, or at least Administrative and Security → choose a destination: a Log Analytics workspace, a storage account or an event hub → Save
Learn more ↗PowerShell
az monitor diagnostic-settings subscription create --subscription 11111111-1111-1111-1111-111111111111 --name ActivityLog-to-LA --location global --workspace <workspace-id> --logs '[{"category":"Administrative","enabled":true},{"category":"Security","enabled":true},{"category":"ServiceHealth","enabled":true},{"category":"Alert","enabled":true},{"category":"Recommendation","enabled":true},{"category":"Policy","enabled":true},{"category":"Autoscale","enabled":true},{"category":"ResourceHealth","enabled":true}]' - 40Should fixAzure
Enable Defender for Cloud plans
Subscription: Contoso Production
What this means
Subscription "Contoso Production" has no paid Defender for Cloud plan (only the free configuration recommendations are active).
What to do
Enable at least Defender for Cloud Servers + Storage at Standard tier — they catch the highest-impact attack patterns (server EDR, ransomware on storage).
- Who does it
- Whoever runs your Azure
- How long
- Not estimated
1 attack path weakenedTechnical detailssetting, steps
The setting
Defender for Cloud workload plans
Microsoft Defender for Cloud → Environment settings → Contoso Production → Defender plans
Now: No paid plan (configuration recommendations only)→Should be: ≥Servers Plan 2 + Storage Standard (catches the highest-impact attack patterns)
Open in the admin portal ↗Learn more ↗Steps
- portal.azure.com → Microsoft Defender for Cloud → Environment settings → Contoso Production → Defender plans
- Enable: Servers Plan 2 (or Plan 1), Storage, SQL servers on machines, App Service
- Save.
- 41Should fixIdentity
Require phishing-resistant MFA for admins
What this means
No Conditional Access policy requires phishing-resistant MFA for admins. They can sign in with a text code or an app approval, which a fake sign-in page can relay.
What to do
Register a passkey or Windows Hello for Business for every admin first, plus two break-glass accounts, then require the built-in "Phishing-resistant MFA" authentication strength for admin roles in Conditional Access.
- Who does it
- Your IT person or provider
- How long
- About 1.5 hours
- Before you change it
- An admin with no passkey, Windows Hello or certificate registered cannot sign in once the policy is on. Keep break-glass accounts excluded and protected separately.
1 attack path weakenedTechnical detailssetting, steps
The setting
Conditional Access: require phishing-resistant MFA for administrators
Entra admin center → Protection → Conditional Access → Create new policy from templates → Require phishing-resistant multifactor authentication for administrators
Now: Admins can satisfy MFA with a method that can be phished→Should be: Admin roles must use the Phishing-resistant MFA strength
Open in the admin portal ↗Learn more ↗Steps
- Entra admin center → Protection → Authentication methods → enable Passkey (FIDO2)
- Have each admin register a passkey or Windows Hello for Business
- Create the template policy in Report-only, check the sign-in logs, then turn it on
- 42Should fixIdentity
Sign out idle web sessions
What this means
No idle session timeout is set. Someone who walks away from a shared or home computer leaves Outlook on the web, SharePoint and the admin centers signed in.
What to do
Turn on Idle session timeout in the Microsoft 365 admin center and set it to 3 hours or less.
- Who does it
- Your IT person or provider
- How long
- About 5 minutes
- Before you change it
- People on shared or unmanaged computers are signed out after the idle period, and it replaces any separate Outlook on the web and SharePoint timeouts. Desktop and mobile apps are not affected, nor are managed devices, single sign-on from a joined device or anyone who chose "Stay signed in".
Technical detailssetting, steps
The setting
Idle session timeout
Microsoft 365 admin center → Settings → Org settings → Security & privacy → Idle session timeout
Now: Off→Should be: On, 3 hours or less
Open in the admin portal ↗Learn more ↗Steps
- admin.microsoft.com → Settings → Org settings → Security & privacy → Idle session timeout (turning it on the first time needs a Global Administrator)
- Turn it on and choose 3 hours or less → Save
- It applies to the whole organisation. To limit it to unmanaged devices, use a Conditional Access policy with app enforced restrictions (needs Entra ID P1)
- 43Should fixIdentity
Review recent privileged changes
What this means
2 privileged changes were made in your tenant. 1 of these granted an app access or added credentials to an app identity — the two changes that let someone back in after a password reset. Every item below is also normal administration; the point is to confirm each one was you or your IT provider. Reviewed 12 days of audit history.
What to do
Work down the list and confirm each change was expected. Anything you do not recognise — especially an app consent or a new app secret — should be revoked before anything else, because it survives a password reset.
- Who does it
- Your IT person or provider
- How long
- About 9 minutes
- Before you change it
- Revoking an app consent stops that app working for everyone until it is re-consented. Confirm what the app is used for before removing it.
Technical detailssetting, steps
The setting
Microsoft Entra audit logs
Entra admin center → Monitoring & health → Audit logs
Open in the admin portal ↗Learn more ↗Steps
- entra.microsoft.com → Monitoring & health → Audit logs
- Filter by the date and activity shown above to see the full record, including the IP address
- For an app consent you do not recognise: Identity → Applications → Enterprise applications → the app → Permissions → Review permissions → revoke
- For an app secret you do not recognise: Identity → Applications → App registrations → the app → Certificates & secrets → delete it
- For an unexpected role addition: Identity → Roles & admins → the role → remove the member
- 44Should fixEmail
Enable Safe Links
What this means
Judged: Built-In Protection Policy. Built-In Protection Policy: AllowClickThrough is true (people can click through to a flagged site). 1 to tighten. Exceptions to Built-in protection, and which people a preset policy covers, are not readable through Microsoft's configuration snapshots.
What to do
Turn on or tighten: AllowClickThrough. Microsoft's Standard and Strict presets set these.
- Who does it
- Your IT person or provider
- How long
- About 10 minutes
Technical detailssetting, steps
The setting
Safe Links policy
Microsoft Defender portal → Email & collaboration → Policies & rules → Threat policies → Safe Links
Now: AllowClickThrough = true→Should be: EnableSafeLinksForEmail/Teams/Office = true; AllowClickThrough = false; ScanUrls = true
Open in the admin portal ↗PowerShell
Connect-ExchangeOnline Get-SafeLinksPolicy | Format-List Name, EnableSafeLinksForEmail, EnableSafeLinksForTeams, EnableSafeLinksForOffice, AllowClickThrough, ScanUrls
- 45Should fixIdentity
Require MFA to join or register devices
What this means
The tenant device-registration MFA switch is off and no enabled Conditional Access policy targets the register-or-join-device user action. Other sign-in protections may still apply; this identifies a gap in dedicated registration MFA controls.
What to do
Review MFA for device registration using either the tenant setting or Conditional Access. Review exclusions and test the impact before enforcement.
- Who does it
- Your IT person or provider
- How long
- About 15 minutes
- Before you change it
- New MFA requirements can affect user-driven enrollment. Validate your enrollment methods and emergency-access procedure before changing enforcement.
Technical detailssetting, steps
The setting
Device registration MFA
Entra admin center → Identity → Devices → Overview → Device settings
Now: notRequired→Should be: Required through the tenant setting or an appropriately scoped Conditional Access policy
Open in the admin portal ↗Learn more ↗Steps
- Open Entra ID → Devices → Overview → Device settings and review the device-registration MFA setting.
- If using Conditional Access, target the 'Register or join devices' user action and require MFA or an MFA authentication strength. Review emergency-access exclusions explicitly.
- For the Conditional Access route, set the tenant device-registration MFA switch to No as Microsoft requires; test in report-only before enabling the CA policy.
- 46Should fixIdentity
Limit what ordinary users can create
What this means
Ordinary users can register applications and create security groups. These are Microsoft's defaults, and an attacker who gets into one ordinary account can do the same.
What to do
Switch off what your users do not need. Administrators and the roles made for the job (Application Developer, for example) keep the ability.
- Who does it
- Your IT person or provider
- How long
- About 15 minutes
- Before you change it
- People who register applications or create security groups today will need an administrator or a role that allows it. Check who has done so recently before switching these off.
Technical detailssetting, steps
The setting
Default user role permissions
Entra admin center → Entra ID → Users → User settings
Now: Users can register applications; Users can create security groups→Should be: None of these left open to ordinary users
Open in the admin portal ↗Learn more ↗Steps
- entra.microsoft.com → Entra ID → Users → User settings → 'Users can register applications' → No
- entra.microsoft.com → Entra ID → Groups → General → 'Users can create security groups in Azure portals, API or PowerShell' → No
- 47Should fixLogging & Alerting
Use a monitored notification mailbox
What this means
Tenant notification contact gaps: No 'Technical contact' email(s) configured — service health incidents and major change announcements go to the tenant's original sign-up contact only; No Security & Compliance notification contact configured — security advisories from Microsoft (including breach notifications) may not reach your security team. Use monitored shared mailboxes ([email protected], alerts@) so notifications survive personnel changes.
What to do
Set the technical and security notification addresses to monitored shared mailboxes (or distribution lists) that survive individual personnel changes. Also: in each admin's personal Message Center preferences, ensure they're subscribed to relevant categories.
- Who does it
- Your IT person or provider
- How long
- About 15 minutes
- Before you change it
- If you change the technical contact to a shared mailbox that no one watches, you've moved the problem rather than fixed it. The shared mailbox needs an owner who reviews it (or it forwards to a SIEM that does).
Technical detailssetting, steps
The setting
Tenant technical & security notification contacts
Microsoft 365 admin center → Settings → Org settings → Organization profile → Data location and contact information
Now: Technical: (none); Security & Compliance: (none)→Should be: Both technical and security/compliance contacts set to monitored shared mailboxes on a corporate domain
Open in the admin portal ↗Steps
- admin.microsoft.com → Settings → Org settings → Organization profile → Data location and contact information
- Set 'Technology contact' to a monitored shared mailbox (e.g. [email protected])
- Set 'Security and Compliance' contact
- Then: admin.microsoft.com → Health → Message center → Edit preferences
- Subscribe a shared mailbox to weekly digest + critical change notifications
- Then (for each admin): admin.microsoft.com → Health → Service health → 'Preferences' → 'Email me when incidents/advisories occur' → On
Learn more ↗PowerShell
Connect-MgGraph Update-MgOrganization -OrganizationId (Get-MgOrganization).Id -TechnicalNotificationMails @('[email protected]') -SecurityComplianceNotificationMails @('[email protected]') - 48Should fixExternal Attack Surface
Certificates expiring within 30 days
What this means
1 certificate(s) expire within 30 days — schedule replacement. Port and service data as observed on 2026-11-09 (Shodan index, not a live probe).
What to do
Schedule renewal.
- Who does it
- Whoever runs your website and domain
- How long
- About 15 minutes
Technical detailssetting
The setting
TLS certificate(s) expiring within 30 days (1)
The hosting environment / load balancer / CDN serving each affected service
Now: 1 certificate(s) expiring within 30 days→Should be: Renewals scheduled or automated well before the 30-day mark
Open in the admin portal ↗ - 49Should fixIdentity
Conditional Access policies stuck in Report-only
What this means
1 policy has sat in Report-only for 60+ days — the oldest for 95 days. Report-only evaluates the policy on every sign-in, writes the verdict to the sign-in log, and then lets the sign-in through anyway. These policies appear in the portal as configured protection and enforce nothing. Either promote them to On or turn them Off, so the policy list reflects what is actually being enforced.
What to do
For each policy: review its Report-only impact, then switch it to On. If the review shows it would break something you are not ready to fix, turn it Off and note why — leaving it in Report-only just disguises the gap.
- Who does it
- Your IT person or provider
- How long
- About 20 minutes
- Before you change it
- Promoting a policy to On starts enforcing it immediately, on the next sign-in. That is the point — but it is also why the Policy impact review comes first. Service accounts and legacy clients are the usual breakage.
Technical detailssetting, steps
The setting
Conditional Access — policy state
Entra admin center → Protection → Conditional Access → Policies → (policy) → Enable policy
Now: 1 policy in Report-only for 60+ days (oldest 95 days)→Should be: Every policy is either On (enforcing) or Off — none parked in Report-only long-term
Open in the admin portal ↗Learn more ↗Steps
- entra.microsoft.com → Protection → Conditional Access → Policies
- Open each policy listed in the evidence → Policy impact tab
- Review which sign-ins would have been blocked or challenged over the last 30 days
- If the impact is acceptable: Enable policy → On → Save
- If it would break something: fix that first, or set Enable policy → Off and record the decision
- Check the Created by column — policies marked 'Microsoft' are managed by Microsoft and normally enable themselves; if one is still Report-only well past 45 days, someone has held it back deliberately
- 50Should fixExternal Attack Surface
Retire outdated TLS versions
What this means
1 public service(s) still offer deprecated SSL/TLS protocols. These are vulnerable to downgrade attacks (an on-path attacker forces the weak version), fail PCI-DSS, and break for modern clients. Disable everything below TLS 1.2. Port and service data as observed on 2026-11-09 (Shodan index, not a live probe).
What to do
Disable SSLv2/SSLv3/TLS 1.0/TLS 1.1 at each service (web server, load balancer, CDN, or appliance) and require TLS 1.2+ (ideally 1.3). Verify with SSL Labs afterward.
- Who does it
- Whoever runs your website and domain
- How long
- About 20 minutes
- Before you change it
- Very old clients (legacy point-of-sale, ancient Android/IE) that only speak TLS 1.0 will stop connecting — which is the point, but confirm you don't depend on any.
Technical detailssetting
The setting
Services offering deprecated TLS/SSL (1)
TLS configuration of each affected web server / load balancer / CDN / appliance
Now: vpn.contoso.com:3389 (TLS 1.0)→Should be: TLS 1.2 minimum (prefer 1.3); SSLv2/SSLv3/TLS 1.0/1.1 disabled
Open in the admin portal ↗ - 51Should fixApps & Consent
Review delegated app consent grants
What this means
1 dangerous delegated scope was granted for individual users. No classified tenant-wide dangerous grant was found.
What to do
Review each grant and revoke any the business does not recognize; then restrict future user consent.
- Who does it
- Your IT person or provider
- How long
- Not estimated
- Before you change it
- Users of a revoked app lose its access until an admin re-consents deliberately.
Technical detailssetting, steps
The setting
Permissions (delegated)
Entra admin center → Enterprise applications → (app) → Permissions
Open in the admin portal ↗Steps
- Entra admin center → Enterprise applications
- Open each flagged app → Permissions
- Revoke the delegated grants you do not recognize
- Tighten user consent under Consent and permissions → User consent settings
- 52Should fixExternal Attack Surface
Self-signed TLS certificates
What this means
2 public-facing service(s) use self-signed certificates. Users get security warnings, and the certificate provides no authentication of identity — making MitM attacks indistinguishable from normal connection. Port and service data as observed on 2026-11-09 (Shodan index, not a live probe).
What to do
Replace with a CA-issued certificate (Let's Encrypt is free for any public hostname).
- Who does it
- Whoever runs your website and domain
- How long
- About 30 minutes
Technical detailssetting
The setting
Self-signed TLS certificate(s) on public services (2)
The hosting environment / load balancer / CDN serving each affected service
Now: 2 service(s) presenting self-signed certificates to the public internet→Should be: All public services present certificates issued by a trusted CA (Let's Encrypt, ACM, commercial CA)
Open in the admin portal ↗ - 53Should fixAzure
Use managed identities
What this means
0% managed-identity coverage on compute (0 of 2). Target ≥80%.
What to do
Enable system-assigned managed identity on each compute resource that authenticates to other Azure services. Removes the need to store credentials in app config / Key Vault references.
- Who does it
- Whoever runs your Azure
- How long
- Not estimated
Technical detailssetting, steps
The setting
Managed identity (Identity blade)
Azure portal → Virtual machines / App Services → [resource] → Identity → System assigned → On
Now: 0% coverage (0/2 resources have managed identity enabled)→Should be: ≥80% of compute resources with system-assigned managed identity enabled
Open in the admin portal ↗Learn more ↗Steps
- portal.azure.com → Virtual machines / App Services → [resource]
- Identity blade → System assigned → On → Save
- Update the resource's app code to use DefaultAzureCredential (idiomatic for both Azure SDKs).
- 54Should fixAzure
Use private endpoints for storage
What this means
0% of storage accounts use private endpoints (0 of 1). Target ≥80%.
What to do
Add private endpoints to storage accounts that hold sensitive data. Then disable public network access on those accounts.
- Who does it
- Whoever runs your Azure
- How long
- Not estimated
Technical detailssetting, steps
The setting
Private endpoint connections
Storage account → Networking → Private endpoint connections
Now: 0% coverage (0/1)→Should be: ≥80% coverage
Open in the admin portal ↗Learn more ↗Steps
- Storage account → Networking → Private endpoint connections → + Private endpoint
- Bind to a subnet in your application's VNet
- After private endpoint is connected: Public network access → Disabled.
- 55Should fixAzure
Require Entra-only sign-in for SQL
Subscription: Contoso Production
What this means
1 SQL Server in subscription "Contoso Production" still accepts SQL logins next to Entra sign-in.
What to do
Move applications and people to Entra sign-in, then turn on Microsoft Entra-only authentication for each server.
- Who does it
- Whoever runs your Azure
- How long
- Not estimated
- Before you change it
- Anything still connecting with a SQL login stops working as soon as this is on, including the server's own SQL admin login. Find those connections first.
Technical detailssetting, steps
The setting
Support only Microsoft Entra authentication for this server
SQL server → Security → Microsoft Entra ID
Now: SQL logins and Entra sign-in→Should be: Entra sign-in only
Open in the admin portal ↗Steps
- SQL server → Security → Microsoft Entra ID
- Confirm an Entra admin is set
- Tick Support only Microsoft Entra authentication for this server → Save
Learn more ↗PowerShell
az sql server ad-only-auth enable --resource-group <rg> --name <server>
- 56Should fixIdentity
Emergency access (break-glass) accounts
What this means
No Global Administrator is excluded from any enabled Conditional Access policy: all 2 standing Global Administrators are inside the same policy set that governs everyone else. A mistyped policy, an expired federation certificate or an MFA provider outage would lock every administrator out of this tenant at once, and recovery then runs through Microsoft support rather than through you. This is an availability risk, not a vulnerability — nothing here is exploitable; the escape hatch is simply missing. Assessed 2 standing Global Administrators against 1 enabled Conditional Access policy; 1 report-only policy was ignored because a report-only policy cannot block a sign-in. We checked excludeUsers, excludeGroups (expanded transitively) and excludeRoles on every enabled policy. Entra ID has no attribute marking an account as break-glass, so this check looks for the configuration Microsoft's emergency-access guidance asks for — a cloud-only account with a standing Global Administrator assignment, explicitly excluded from every enabled Conditional Access policy that would otherwise apply to it, and with no interactive sign-in in the last 90 days. An account whose sign-in activity could not be read is not counted. Weaker signals (no licences, no mailbox, naming convention) are shown per account but do not decide the result, because a forgotten former-employee admin looks identical on those.
What to do
Create two cloud-only Global Administrator accounts, exclude both from every enabled Conditional Access policy, give each its own passkey or certificate held by a different person, and alert on any sign-in by them.
- Who does it
- Your IT person or provider
- How long
- About 45 minutes
- Before you change it
- Excluding an account from Conditional Access means MFA and device-compliance rules stop applying to it — that is the point, and it is why the account must sign in with its own phishing-resistant method and be monitored. Expect these accounts to surface in other findings too: they are dormant by design (stale admin accounts) and often have passwords set never to expire, which the password-expiry check flags. Those are known, intentional collisions, not contradictions.
Technical detailssetting, steps
The setting
Conditional Access — policy user exclusions
Entra admin center → Protection → Conditional Access → Policies
Now: 0 of 2 standing Global Administrators are excluded from any enabled policy→Should be: 2 cloud-only emergency access accounts excluded from every enabled Conditional Access policy
Open in the admin portal ↗Learn more ↗Steps
- entra.microsoft.com → Identity → Users → New user: create two cloud-only accounts on the tenant's *.onmicrosoft.com domain
- Assign each one Global Administrator permanently — NOT PIM-eligible, since an eligible role cannot be activated if PIM or Conditional Access is what has broken
- Register a passkey (FIDO2 security key) on each, or certificate-based authentication if you already run a PKI, using a method your other admin accounts do not use; Microsoft's mandatory MFA for admin portals applies to these accounts too, so a password alone will not get in
- Store each account's key with a different person, in a separate secure location
- Protection → Conditional Access → open every enabled policy → Users → Exclude → add both accounts (a dedicated 'Emergency access' group is easier to keep in sync than per-policy user exclusions)
- Set an alert on sign-ins by these accounts (Entra sign-in logs → alert rule) so a use is never silent
- Test both accounts on a schedule — Microsoft suggests every 90 days, and after every Conditional Access change
- 57Should fixEmail
MTA-STS not published
What this means
contoso.com receives mail but has no MTA-STS policy. Sending servers fall back to opportunistic TLS, which an on-path attacker can strip (STARTTLS downgrade) to read or alter inbound email in transit — including password resets and invoices.
What to do
Publish an MTA-STS policy: a TXT record at _mta-sts.<domain> plus a policy file at https://mta-sts.<domain>/.well-known/mta-sts.txt (start in mode=testing, then enforce). Microsoft 365 supports MTA-STS for hosted domains.
- Who does it
- Your IT person or provider
- How long
- About 45 minutes
- Before you change it
- Enforcing MTA-STS before your MX/TLS certs are correct can defer legitimate mail. Publish mode=testing first and watch TLS-RPT reports.
Technical detailssetting
Learn more ↗The setting
MTA-STS for contoso.com
DNS: TXT at _mta-sts.contoso.com + an mta-sts.<domain> HTTPS policy endpoint
Now: (no MTA-STS policy)→Should be: v=STSv1; id=<timestamp> + policy at mode=enforce
Open in the admin portal ↗ - 58Should fixIdentity
Find sign-ins where no second factor was required
What this means
3 successful sign-ins by 2 accounts in the last 30 days required only one factor. Microsoft did not require a second factor. The person may still have used Windows Hello or a passkey, and a policy may have relied on a managed device, so check why before changing anything.
What to do
Open a few of these sign-ins and read the Authentication details and Conditional Access tabs to see what was used and why one factor was enough. If no policy meant to allow it, require MFA for all users and all cloud apps.
- Who does it
- Your IT person or provider
- How long
- About 1 hour
- Before you change it
- A new MFA policy prompts everyone it covers, including accounts that have not registered a method. Register people first and keep an emergency access account excluded.
Technical detailssetting, steps
The setting
Conditional Access: require MFA
Entra admin center → Monitoring & health → Sign-in logs → filter Authentication requirement = Single-factor authentication
Now: 3 successful sign-ins requiring one factor in the last 30 days, 2 accounts, 0 admins→Should be: Single-factor sign-ins only where a policy allows them on purpose
Open in the admin portal ↗Learn more ↗Steps
- Entra admin center → Monitoring & health → Sign-in logs
- Add the filter Authentication requirement = Single-factor authentication, and open the sign-ins below
- On each, read the Conditional Access tab, then close the gap with a policy in Report-only first
- 59Should fixIdentity
Admin accounts with mailboxes
What this means
3 privileged accounts have a mailbox. Best practice is dedicated cloud-only admin accounts with NO mailbox — that way phishing the admin requires guessing a UPN that's not in any directory, never appears as a sender, and is never on a public CC.
What to do
Create dedicated admin accounts (e.g. `[email protected]`) without mailbox licenses. Move privileged role assignments to the new accounts; demote or remove privilege from the everyday accounts.
- Who does it
- Your IT person or provider
- How long
- About 1 hour
- Before you change it
- If the admin uses Outlook to manage email AND also clicks 'open admin portal', they'll now need to sign in with a different account. Workflow change.
Technical detailssetting, steps
The setting
Privileged users — mailbox attached
Entra admin center → Identity → Users → All users (review listed admin accounts)
Now: 3 privileged accounts with a mailbox attached→Should be: Privileged accounts are dedicated cloud-only identities with no mailbox license
Open in the admin portal ↗Learn more ↗Steps
- Create a new user without an M365 mailbox license: admin.<name>@<yourdomain>
- Assign the same privileged roles to the new account
- Remove the privileged roles from the existing mailbox-attached account
- Add the new admin account to your tenant-wide MFA Conditional Access policy
- 60MinorAzure
Encrypt VM disks
Subscription: Contoso Production
What this means
Encryption at host is off on 1 virtual machine in subscription "Contoso Production", so its temporary disk and caches are not encrypted. Azure already encrypts its managed disks.
What to do
Enable encryption-at-host on each VM (preferred — requires VM stop). For VMs that can't stop, fall back to Azure Disk Encryption (BitLocker / dm-crypt), which Microsoft plans to retire in September 2028.
- Who does it
- Whoever runs your Azure
- How long
- Not estimated
1 attack path weakenedTechnical detailssetting, steps
The setting
Encryption at host
Virtual machine → Disks → Additional settings (or per-disk Encryption settings)
Now: Off. Azure encrypts the managed disks; temporary disks and caches are not encrypted→Should be: On, or Azure Disk Encryption
Open in the admin portal ↗Steps
- Stop the VM (required for encryption-at-host)
- portal.azure.com → Virtual machines → [VM] → Disks → Additional settings
- Encryption at host → Enabled → Save → Start VM
Learn more ↗PowerShell
az vm update --name <vm> --resource-group <rg> --set securityProfile.encryptionAtHost=true
- 61MinorIdentity
Do not expire passwords
What this means
Password expiry is not set on 2 of 2 domains. Microsoft documents that a password is then valid for 90 days. Setting it to never expire removes the doubt.
What to do
Set passwords to never expire, and rely on MFA and on changing a password when it is known to be exposed.
- Who does it
- Your IT person or provider
- How long
- About 5 minutes
- Before you change it
- Nothing breaks. Check that MFA is in place first: with no expiry, MFA is what limits the use of a stolen password.
Technical detailssetting, steps
The setting
Password expiration policy
Microsoft 365 admin center → Settings → Org settings → Security & privacy → Password expiration policy
Now: Not set (Microsoft documents 90 days)→Should be: Set passwords to never expire (recommended)
Open in the admin portal ↗Steps
- admin.microsoft.com → Settings → Org settings → Security & privacy
- Password expiration policy → tick 'Set passwords to never expire (recommended)'
- Save
- The setting is held on each domain. If a domain named here is still listed after the next scan, set it with the PowerShell command in this finding.
Learn more ↗PowerShell
# Microsoft Graph PowerShell, signed in with Domain.ReadWrite.All Update-MgDomain -DomainId contoso.com -PasswordValidityPeriodInDays 2147483647 Update-MgDomain -DomainId contoso.onmicrosoft.com -PasswordValidityPeriodInDays 2147483647
- 62MinorIdentity
Give every Global Administrator an outside contact
What this means
1 of 2 enabled Global Administrators has no contact route outside the tenant: no alternate email on another domain and no phone number. If the tenant is locked or breached, your own team has no way to reach that admin from the directory.
What to do
Add an alternate email on a different domain and a phone number to each Global Administrator's profile, and have them register the same details as recovery methods.
- Who does it
- Your IT person or provider
- How long
- About 5 minutes
- Before you change it
- Nothing, though other people in the directory can see contact fields, so use details that are fit for work.
Technical detailssetting, steps
The setting
User profile: contact information
Entra admin center → Identity → Users → All users → (admin) → Properties → Contact information
Now: 1 Global Administrator without an outside contact→Should be: Every Global Administrator has an alternate email on another domain or a phone number
Open in the admin portal ↗Learn more ↗Steps
- entra.microsoft.com → Identity → Users → All users → open the admin → Properties → Edit properties
- Contact information → Other emails: add an address outside this tenant; Mobile phone: add a number
- Ask the admin to register the same email and phone at aka.ms/mysecurityinfo
- 63MinorSharing & Collaboration
Block guest re-sharing
What this means
Guests may pass on items shared with them, but only to guests who already exist in your directory (you have none).
What to do
Disable guest re-sharing tenant-wide. Guests can still ask the document owner to invite additional collaborators directly.
- Who does it
- Your IT person or provider
- How long
- About 5 minutes
- Before you change it
- Workflows where a guest forwards a shared document to another contact at their own org will break. They'll need to request an additional direct invite from the document owner.
Technical detailssetting, steps
The setting
SharePoint guest re-sharing
SharePoint admin center → Policies → Sharing → More external sharing settings → 'Allow guests to share items they don't own'
Now: On (guests can re-share)→Should be: Off
Open in the admin portal ↗Steps
- admin.microsoft.com → SharePoint admin center → Policies → Sharing
- Under 'External sharing' → 'More external sharing settings'
- 'Allow guests to share items they don't own' → Off
- Save
Learn more ↗Microsoft Graph
PATCH /beta/admin/sharepoint/settings { "isResharingByExternalUsersEnabled": false } - 64MinorIdentity
Keep password lockout and banned passwords tight
What this means
Password protection could be tighter: the custom banned password list is empty. Password spraying tries a few likely passwords against many accounts, and these two controls decide how much each guess costs.
What to do
Keep the lockout threshold at 10 or lower and the duration at 60 seconds or more. With Entra ID P1, enforce a custom list of your company, product and place names.
- Who does it
- Your IT person or provider
- How long
- About 10 minutes
- Before you change it
- A lower threshold locks out people who mistype repeatedly, though smart lockout tells familiar locations from unfamiliar ones. A new banned term applies the next time someone changes a password, and an update can take several hours to reach everyone.
Technical detailssetting, steps
The setting
Password protection
Entra admin center → Protection → Authentication methods → Password protection
Now: the custom banned password list is empty→Should be: Threshold 10 or lower, duration 60 seconds or more, custom list filled in and enforced
Open in the admin portal ↗Learn more ↗Steps
- entra.microsoft.com → Protection → Authentication methods → Password protection
- Lockout threshold → 10 or lower; Lockout duration in seconds → 60 or higher
- Enforce custom list → Yes, then add your company name, products and city
- 65MinorEmail
TLS-RPT not published
What this means
contoso.com has no TLS-RPT record, so you get no reports when senders fail to deliver over TLS — you'd be blind to an active downgrade or a broken MX certificate.
What to do
Add a TLS-RPT TXT record so you receive aggregate TLS-failure reports.
- Who does it
- Your IT person or provider
- How long
- About 10 minutes
Technical detailssetting
The setting
TLS-RPT for contoso.com
DNS: TXT at _smtp._tls.contoso.com
Now: (no TLS-RPT record)→Should be: v=TLSRPTv1; rua=mailto:[email protected]
Open in the admin portal ↗ - 66MinorIdentity
Review local administrator defaults for device join
What this means
Review the local administrator defaults for device join: the join policy adds all Global Administrators as local administrators; the join policy adds all joining users as local administrators. Confirm these assignments match your device-management needs; this policy read does not establish current local group membership.
What to do
Turn off 'Global administrator role is added as local administrator', make registering users standard users (grant local admin through Intune or the 'Microsoft Entra Joined Device Local Administrator' role instead), and scope who may join devices if you use Autopilot.
- Who does it
- Your IT person or provider
- How long
- About 15 minutes
- Before you change it
- Changing join defaults may affect future enrollment and software installation. Review existing device local groups separately and grant necessary access deliberately through device management.
Technical detailssetting, steps
The setting
Local administrator settings for Microsoft Entra joined devices
Entra admin center → Identity → Devices → Overview → Device settings → Manage Additional local administrators on all Microsoft Entra joined devices
Now: the join policy adds all Global Administrators as local administrators; the join policy adds all joining users as local administrators→Should be: Global administrator role not added as local admin; registering user not local admin (standard user); join limited to the users or groups that need it
Open in the admin portal ↗Learn more ↗Steps
- entra.microsoft.com → Identity → Devices → Overview → Device settings
- 'Global administrator role is added as local administrator on the device during Microsoft Entra join' → No
- 'Registering user is added as local administrator on the device during Microsoft Entra join' → None (or Selected)
- 'Users may join devices to Microsoft Entra' → Selected, if you enrol through Autopilot or IT
- 67MinorIdentity
Decide who can create Microsoft 365 Groups
What this means
Every user can create Microsoft 365 Groups. This is Microsoft's default, and it is a choice rather than a mistake: each group is a Team, a SharePoint site and a mailbox that nobody reviews when it is made.
What to do
If you want a review step, limit creation to a named group. If people making their own Teams suits you, keep the default and review groups some other way, such as expiry or access reviews.
- Who does it
- Your IT person or provider
- How long
- About 20 minutes
- Before you change it
- People outside the named group can no longer create Teams or Planner plans themselves and have to ask. Microsoft's guidance is that the people who keep the ability need Entra ID P1.
Technical detailssetting, steps
The setting
Group.Unified directory settings
Entra admin center → Groups → General
Now: every user can create Microsoft 365 Groups→Should be: Guests cannot own groups; creation limited to a named group if you want a review step
Open in the admin portal ↗Steps
- entra.microsoft.com → Groups → General → 'Users can create Microsoft 365 groups in Azure portals, API or PowerShell' → No
- Name a security group of approved creators in GroupCreationAllowedGroupId (PowerShell, see the link below)
- Set AllowGuestsToBeGroupOwner to false (PowerShell)
Learn more ↗PowerShell
Connect-MgGraph -Scopes Directory.ReadWrite.All $s = Get-MgGroupSetting | Where-Object { $_.DisplayName -eq 'Group.Unified' } # Review $s.Values, then update EnableGroupCreation, GroupCreationAllowedGroupId and AllowGuestsToBeGroupOwner - 68MinorIdentity
Finish the authentication methods migration
What this means
The move to stronger sign-in methods is unfinished: the registration campaign is off.
What to do
Set the registration campaign to Microsoft managed or Enabled, for all users.
- Who does it
- Your IT person or provider
- How long
- About 20 minutes
- Before you change it
- Methods enabled only in the legacy MFA or SSPR policy stop working at Migration Complete, so check the modern policy first. The campaign prompts people at sign-in until they register or snooze.
Technical detailssetting, steps
The setting
Authentication methods policy: Manage migration and Registration campaign
Entra admin center → Protection → Authentication methods → Policies
Now: the registration campaign is off→Should be: Migration Complete; registration campaign on for all users
Open in the admin portal ↗Learn more ↗Steps
- entra.microsoft.com → Protection → Authentication methods → Policies: confirm every method people rely on is enabled here
- Manage migration → Migration Complete
- Authentication methods → Registration campaign → State: Microsoft managed or Enabled → Include: All users → Save
- 69MinorIdentity
Review the tenant setting for Windows LAPS backups
What this means
Microsoft Entra password backup for Windows LAPS is disabled. Review whether you manage supported Windows devices and use Entra for backups. This does not establish shared passwords, missing rotation, or the absence of another password-management solution.
What to do
If using Microsoft Entra for Windows LAPS backups, enable the tenant setting and deploy a client policy to rotate and back up local administrator passwords.
- Who does it
- Your IT person or provider
- How long
- About 30 minutes
- Before you change it
- Applies to supported Entra-joined or hybrid-joined Windows devices. Review existing password-management tools and pilot the client policy before deployment.
Technical detailssetting, steps
The setting
Enable Microsoft Entra Local Administrator Password Solution (LAPS)
Entra admin center → Identity → Devices → Overview → Device settings
Now: No→Should be: Yes, with an Intune LAPS policy assigned to Windows devices
Open in the admin portal ↗Learn more ↗Steps
- entra.microsoft.com → Identity → Devices → Overview → Device settings → 'Enable Microsoft Entra Local Administrator Password Solution (LAPS)' → Yes
- intune.microsoft.com → Endpoint security → Account protection → Create policy → Windows LAPS
- Set the backup directory to Microsoft Entra ID, choose a password rotation policy and assign it to supported Windows devices; verify successful backups.
- 70MinorApps & Consent
Expired app credentials still present
What this means
1 expired credential(s) are still attached to app registrations. Cleanup hygiene — they can't be used but they clutter your credential inventory.
What to do
Delete the expired credential entries via the portal or Graph API.
- Who does it
- Your IT person or provider
- How long
- Not estimated
Technical detailssetting
The setting
App registration credentials
Entra admin center → Identity → Applications → App registrations → (each app) → Certificates & secrets
Now: 1 expired credential(s) still attached→Should be: Expired credentials removed
Open in the admin portal ↗ - 71MinorExternal Attack Surface
DNSSEC not enabled
What this means
contoso.com does not publish DNSSEC (no DS record in the parent zone). Without DNSSEC, an attacker who can poison a resolver's cache or sit on the network path can forge answers for your domain — redirecting mail (MX) or web traffic (A) to infrastructure they control.
What to do
Enable DNSSEC at your DNS provider / registrar. Most (Cloudflare, Route 53, Azure DNS) enable it with one click, then publish the DS record to your registrar.
- Who does it
- Whoever runs your website and domain
- How long
- About 30 minutes
- Before you change it
- A botched DNSSEC rollout (mismatched DS after a key rotation) can make the domain unresolvable. Follow the provider's guided flow and verify before/after.
Technical detailssetting
Learn more ↗The setting
DNSSEC for contoso.com
Your DNS provider → DNSSEC (enable), then your registrar → add the DS record
Now: DNSSEC not enabled (no DS record)→Should be: DNSSEC enabled and DS record published at the registrar
Open in the admin portal ↗ - 72MinorAzure
Require TLS 1.2 on storage
Subscription: Contoso Production
What this means
1 storage account in subscription "Contoso Production" still has its minimum TLS version set below 1.2. Azure Storage stopped accepting TLS 1.0 and 1.1 on 3 February 2026, so the setting no longer lets older connections in. Set it to 1.2 so it matches.
What to do
Set minimumTlsVersion to TLS1_2 on each storage account.
- Who does it
- Whoever runs your Azure
- How long
- Not estimated
Technical detailssetting, steps
The setting
Minimum TLS version
Storage account → Configuration → Minimum TLS version
Now: TLS1_0→Should be: TLS1_2
Open in the admin portal ↗Learn more ↗PowerShell
az storage account update --name <name> --resource-group <rg> --min-tls-version TLS1_2
- 73MinorAzure
Turn off storage shared key access
Subscription: Contoso Production
What this means
1 storage account in subscription "Contoso Production" allows shared key access, because it has no value set and Microsoft's default is to allow it.
What to do
Find out what still signs in with an account key or a SAS token signed by one, move it to Entra, then turn shared key access off.
- Who does it
- Whoever runs your Azure
- How long
- Not estimated
- Before you change it
- Anything that signs in with an account key or a SAS signed by one gets a 403. Check first for Azure Files without Entra permissions, Azure Cloud Shell file shares, Storage Explorer on file shares, and any service that connects with a connection string.
Technical detailssetting, steps
The setting
Allow storage account key access
Storage account → Settings → Configuration
Now: Allowed→Should be: Disabled
Open in the admin portal ↗Learn more ↗Steps
- Storage account → Monitoring → Metrics: split Transactions by Authentication to see what uses shared keys
- Move those applications to Entra sign-in or a user delegation SAS
- Storage account → Settings → Configuration → Allow storage account key access → Disabled → Save
Bigger lifts: 1 fix
Fixes that take longer or wait on an earlier fix. Some need other teams or a planned change window.
- Effort
- About 3 hr
- Projected
- 100
- Lift
- +54
- 01Should fixExternal Attack Surface
Patch internet-facing services
What this means
3 CVE match(es) across the internet-facing footprint — 3 not scored. No severity score was found for any of them; that does not make them minor. Each is inferred from software a public index saw on the address, not tested against the service. Port and service data as observed on 2026-11-09 (Shodan index, not a live probe).
What to do
Patch or upgrade each affected service. For products you can't immediately patch (legacy appliances, vendor-managed software), restrict the port at the firewall and pursue vendor support.
- Who does it
- Whoever runs your website and domain
- How long
- About 3 hours
- Before you change it
- Patches and upgrades can introduce compatibility regressions. Apply in a non-production environment first if possible.
1 attack path weakenedTechnical detailssetting, steps
The setting
Internet-facing services with known CVEs (top: CVE-2019-0708 on vpn.contoso.com)
The hosting environment for each affected service (cloud VM, on-prem server, appliance admin panel) — firewall as a stop-gap
Now: 3 CVE match(es) (0 critical, 0 high, 0 medium), 3 not scored→Should be: All internet-facing services patched above the affected version range, or restricted to known-IP allowlists
Open in the admin portal ↗Learn more ↗Steps
- For each CVE: read the NVD entry (linked) to confirm applicability
- Identify the device hosting the service (cloud VM, on-prem server, appliance)
- Patch to a version above the affected range, OR
- Restrict the service to known-IP allow-lists at the network edge until patch is available
- Subscribe to vendor security advisories so future CVEs are caught faster
Modeled effect on attack paths
These are projections from the current findings, not verified outcomes.
Projected to close
21
- Domain spoofing → wire transfer / invoice fraud
- Endpoint compromise → OneDrive sync fanout → org-wide ransomware
- Azure account without MFA → subscription takeover
- Detection and logging gaps → an attacker goes unnoticed
- Internet-exposed admin surface → credential reuse → tenant takeover
- Known CVE on public service → web shell → ransomware deployment
- Password spray → mailbox takeover → business email compromise
- Lookalike domain set up for email → mail posing as you or a supplier → invoice fraud
- Fake sign-in page relays an admin's MFA → stolen session → weeks of admin access
- Database open to the internet → passwords or leaked keys tried from anywhere → data copied out
- Internet-exposed Azure VM → lateral movement → ransomware
- Public Azure SQL endpoint → data exfiltration
- Leaked service-principal secret → subscription-scope cloud takeover
- Illicit OAuth consent → silent mailbox exfiltration
- Teams DM phishing → credential / OAuth-grant compromise
- Account with no MFA → attacker registers their own MFA method → lasting access
- Laptop lost or stolen → disk read without a password → company files exposed
- Device code phishing → person completes MFA for the attacker → mailbox and files
- Leaked or added app credential → signs in as the app → data access with no user or MFA
- Key Vault open to any network → a stolen sign-in reaches it from anywhere → keys deleted for good
- Azure storage configuration → possible unauthorized access
Projected to remain
0