Active Directory security assessment Know exactly how exposed each directory is. Then fix it, one reviewed change at a time.
A read-only security assessment evaluates your Active Directory domains, Entra ID tenants and Okta organizations against 144 exposure indicators and the DISA STIG benchmarks, and turns the findings into a health score you can track quarter over quarter.
- 144 exposure indicators: 64 for Active Directory, 43 for Entra ID, 37 for Okta
- A 0 to 100 health score per directory, with a trend across every run
- Guided fixes for AD and Entra ID, shown change by change before they are applied
One number per directory, and the five categories behind it
There are four levels of exposure items: Critical, High, Medium or Low.
A domain with excellent hygiene but a critical privileged-access problem shows as Poor, not Good, so a single serious weakness never hides behind strong averages.
Scores fall into four bands, and the dashboard card shows the band, the number and the trend for each domain, tenant and organization at a glance.
- Good
- 90 and above Very little is open, and no category is lagging far behind the others. Keep an eye on the trend.
- Fair
- 70 to 89 Sound overall, with a number of findings still open in one or more categories. Work through them as routine.
- Poor
- 40 to 69 Many findings are open, or one category is weak enough to hold the whole score down. Plan the fixes now, weakest category first.
- Critical
- Below 40 Serious findings are open across most categories, not just one. Treat it as urgent and start with the critical indicators.

-
Privileged Access Who holds administrative rights, how they got them, and whether they should
-
Credentials and Kerberos Weak, stale, reversible and exposed credentials, and the accounts that invite attack
-
Delegation, ACLs and Applications Delegation, permissions and certificate templates in AD; application permissions and consent in Entra ID and Okta
-
Infrastructure and Hygiene How each domain, tenant and organization is configured, from protocols to policies
-
Indicators of Compromise Signs an attacker may already be inside, from touched decoys to suspicious credentials
Written for each directory, not translated from one to the others
Active Directory, Entra ID and Okta fail in different ways. Each has its own set of indicators, and each indicator carries a stable ID so a finding means the same thing from one release to the next.
Active Directory
- 13 privileged access
- 16 credentials and Kerberos
- 13 delegation, ACLs and certificate templates
- 18 infrastructure and hygiene
- 4 indicators of compromise
Unconstrained delegation, Kerberoastable service accounts, dangerous permissions on Tier 0 objects, passwords in Group Policy preferences, SMBv1 on domain controllers, stale computers, and more.
Entra ID
- 11 privileged access
- 10 credentials
- 8 applications and consent
- 10 infrastructure
- 4 indicators of compromise
Permanent Global Administrators, administrators without MFA, applications with takeover-grade Graph permissions, risky users, partner relationships, Conditional Access gaps, and more.
Okta
- 9 privileged access
- 11 credentials
- 3 applications
- 10 infrastructure
- 4 indicators of compromise
Super Administrators and their factors, API tokens without network zones, long sessions, ThreatInsight and support access settings, lockout and password policies, and more.
Alongside the DISA STIG benchmarks
The same run evaluates the published Security Technical Implementation Guides where they apply, so a compliance finding and an exposure finding sit side by side in one report.
- Active Directory Domain
- Windows Server 2016, 2019, 2022 and 2025
- Microsoft Entra ID
- Okta Identity as a Service
Read-only by design
The analysis only reads. It binds to a domain controller, queries Microsoft Graph or the Okta management API with the account you give it, and writes nothing back. The only time Identity CORE changes a directory is a guided fix you have reviewed and approved, or a rollback you asked for.
Every finding names the objects behind it
An indicator is not a yes or no. It lists the accounts, groups, computers, applications or policies it found, with the evidence it used, so the person fixing it starts with a list and not a search.
Filter the board by severity or category, open an object straight from the finding, and see the same object in the Identity browser or the Auditor without leaving the console.
- Severity, points and the category each indicator counts toward
- The affected objects, with the full count even when the list is long
- Evidence the indicator used, so a finding can be checked by hand
- Deep links from the dashboard and the notification email to the exact indicator
Resolve an indicator without writing the script yourself
Most Active Directory and Entra ID indicators come with a fix. Identity CORE works out the exact changes from the live directory, shows them to you, and applies only what you tick.
- 1
Choose Resolve on an open indicator Only indicators the product knows how to fix offer it. Administrators and managers can run one.
- 2
Review the plan Identity CORE reads the live directory and lists every change: the object, the attribute, its current value and the new one, with a reason and any warning. Changes that need a judgement call come unticked.
- 3
Approve and apply Only the ticked changes are written, each object re-read first so a change that is no longer needed is skipped. The run is recorded with its request and result.
- 4
Confirm or undo The next analysis confirms the fix or reopens it. Every applied change keeps its previous value, and a run can be undone from the activity record.
Show the trend, not just today's number
Run the analysis on a schedule and every run is stored. The trend line shows the score moving as fixes land, and any earlier run can be opened in place of the latest to see what was open then.
Each finding carries a workflow state your team sets: in progress, remediated or ignored. The next run confirms a fix or reopens it if the setting drifts back. Ignored findings are left out of the score, so an accepted risk stops dragging the number down.
For requirements that cannot be checked automatically, an attestation records that someone reviewed it by hand, who and when.
Tell it what is deliberate, so it reports what is not
Every directory has exceptions made on purpose. The analysis configuration lets you declare them, and the indicators treat them accordingly instead of reporting the same accepted item every run.
-
Honey accounts Decoy accounts you deployed to catch credential attacks. One indicator verifies each decoy is still convincing, and reports a touched decoy as an incident.
-
Break-glass administrator The one emergency account kept out of Protected Users on purpose, reported as excluded rather than as a finding.
-
Service accounts Left out of the checks that expect a person behind the account, and counted in the evidence instead.
-
Additional privileged groups Your own groups that should be treated as administrators by the privileged-access and permission indicators.
-
Public-facing servers and their admin groups Feeds the requirement that internet-facing servers are administered separately.
-
Stale computer threshold How many days without a logon before a computer counts as no longer deployed. 90 by default.
Risk views for any account, and an inventory with every run
User, Group and Computer Risks
A card per risk category with its count, and the accounts or machines in the one you pick. Investigate any account from here, run a task script against it, or open it in the Identity browser.
Directory inventory with every run
Users enabled, disabled, inactive and never logged on; password age buckets; smart card and locked-out counts; computers by operating system with stale and domain controller counts; groups, OUs, GPOs, trusts, sites and functional levels. For a tenant: members, guests, synchronized users, MFA registration, role holders, applications, devices and Conditional Access policies.
Scheduled, with a notification that says what matters
Run the analysis nightly or weekly. The notification quotes the score, the inventory and the costliest open indicators, with a link to each.
Common Questions
Does the analysis change anything in my directory?
No. The scan is read-only. The only writes Identity CORE ever makes to a directory are guided fixes you reviewed and approved, and rollbacks you asked for. Both are recorded with who ran them and what changed.
What permissions does it need?
For Active Directory, a read-only account that can bind to a domain controller and read SYSVOL. For Entra ID, an app registration with read permissions on Microsoft Graph; a refused request names the permission to grant. For Okta, a read-only API token. Guided fixes need write permission and ask for it separately.
How long does a run take?
It depends on the size of the directory. Objects are cached once per run and the indicators share them, so even the delegation and permission checks finish in minutes on most domains. Runs are scheduled jobs, so they happen off-hours if you prefer.
Can I export the results for an auditor?
Yes. The Security Posture reports list the open checks of the latest analysis for each domain, tenant and organization with its score, and export to PDF, Excel or CSV like every other report.
See the score for your own directories
Book a demo and we will run the analysis against a lab domain, or walk through yours.