Blog

One Account, Five Hops: How an Ownerless Service Account Reaches Your Finance Database

Rita Gurevich
EmailTwitterLinkedIn

 

Why it matters

An ownerless service account can carry far more privilege than its name or direct memberships suggest. Nested groups, server administrator rights, and service dependencies can connect it to sensitive business systems. Reviewing each system separately can leave those relationships unexplained.

Bringing together account discovery, ownership context, effective access, and service dependencies helps teams identify exposure, determine who should review it, and plan changes with an understanding of the operational impact.

Consider this illustrative example of the full chain:

  • An ownerless domain account belongs to an application group.
  • Nested membership grants it local administrator rights on a server.
  • A second account runs a service on that server and has elevated access to a finance database.
  • If an attacker compromises the first account and obtains usable credentials for the second, the exposure can extend to the database.

A single ownerless service account can reach a production finance database in five hops, and no single tool shows the path. The five hops below show why the important questions are both “What can this account reach?” and “Who is accountable for keeping that access appropriate?”

Hop 1: The ownerless service account

svc-finance-app is an Active Directory domain account created to run an application without a person logged in.

  • Its password is set as “password never expires”.
  • It has no recorded owner. The original administrator who created it has left, and the account is
    absent from regular access reviews.
  • It falls outside access certifications, because reviews are built around people and managers, and
    a service account has no manager.
  • Nobody is watching what it can reach.

Hop 2: The group name doesn’t dictate access

The svc-finance-app account is a direct member of a group called App-Admins. The group name sounds familiar, but it says little about where the group has access.

  • The name suggests administrators of an application; but a group name is often just a label someone typed, not a fact.
  • What a group, and the group’s members, can actually do comes from where it has been granted rights: local admin groups on servers, cloud infrastructure access, file share permissions, database roles.
  • The only way to know is to correlate the group against where it has been granted access, directly and indirectly.
  • The group may have an Owner, but most often the Owner is reviewing only the direct memberships, without context of what the group and its members have access to:
  • Note: the Owner easily could approve App-Admins without realizing that it contributes to administrator access on a server.

Hop 3: Group nesting grants server admin access

So, svc-finance-app is a member of App-Admins … and App-Admins is a member of FinApp-03 … and FinApp-03 has local standing administrative access to server WIN-APP-03.

Through three levels of nesting, the service account becomes a local administrator on WIN-APP-03. Result: svc-finance-app has full admin rights on that server.

Why it goes unnoticed:

  • The server’s admin group shows only FinApp-03.
  • The account’s profile shows only as part of App-Admins.
  • You only see the connection by resolving the entire chain.

Hop 4: The server holds another account’s password

This step creates an opportunity for credential theft: an attacker who gains sufficient control of the server may be able to obtain credentials for another domain account!

How it works:

  • WIN-APP-03 runs a Windows service, such as the finance app’s batch processor, configured to run as a second domain account, svc-finance-db.
  • To start that service automatically, Windows stores svc-finance-db’s password on the server in its registry. This is default Windows behavior, not a misconfiguration.
  • That storage protects the password from regular users, not from local admins. An admin can elevate to SYSTEM and read it with freely available tools anybody (or ‘anything’) can use.
  • While the service is running, its credentials are also in memory, where an admin can retrieve them.
  • MITRE ATT&CK lists this as a standard credential access technique.

Why attackers target this hop:

  • It is a standard lateral movement step: get admin on one server, harvest stored credentials, use them to reach the next system.
  • Servers where privileged domain accounts run services are effectively credential stores.

How common it is:

  • Very common, especially on older application servers. Apps that reach a remote database or file share typically need a domain service account.
  • Group Managed Service Accounts (gMSAs), which rotate passwords automatically, have uneven adoption, and many legacy apps don’t support them.
  • Rotating a stored password risks breaking the app, so passwords stay unchanged for years.

Where this hop does not apply: services running as LocalSystem, NetworkService, or virtual accounts store no domain password. gMSAs shorten the exposure window but do not eliminate it.

Hop 5: That second account has elevated access to the finance database

The harvested account, svc-finance-db, has admin rights on the production finance database.

  • svc-finance-db connects to PROD-SQL through Windows authentication.
  • It has been granted db_owner rights on the finance database, because “the app needed it” and nobody scoped it down.
  • db_owner provides broad control within that database.
  • Result: whoever (or ‘whatever’) controls the ownerless account from Hop 1 can control the production finance database.

If an attacker obtains usable credentials for svc-finance-db and can connect and authenticate to the SQL Server, the attacker could exercise the account’s database privileges.

Important: The first account’s exposure therefore warrants review alongside the second account’s service dependencies and database access.

Why the path stays invisible

Each hop is owned by a different team, and each is legitimate on its own. Infrastructure hardening often follows technology boundaries, with each team responsible for its own controls.

  • The AD team owns the groups and nested membership.
  • The server team manages local administrator rights and service configuration.
  • The DBAs manage database logins and roles.
  • IGA coordinates access reviews, while PAM teams manage privileged credentials.
  • No one sees the chain end to end, so no one sees the risk.

Each team may validate its own configuration without understanding how access connects across systems. A nested group membership, a service dependency, and an elevated database role can collectively create exposure that separate reviews overlook.

What SPHEREboard and the Identity Map reveal

SPHEREboard identifies the access and dependencies behind each step, and the Identity Map brings them together to show how far an account can reach.

  • Effective group membership, including nesting, resolved down to the account.
  • Who effectively holds local admin rights on each server.
  • Which domain accounts run services on each server.
  • Which accounts hold privileged roles on databases such as SQL Server and Oracle.
  • Put together, that is the exposure an attacker is hunting for, shown before they find it.

What to do about it:

  • Assign an owner to every service account.
  • Remove stale nested groups from server admin groups.
  • Reduce standing administrator access and apply just-in-time access to administrative tasks where supported
  • Scope database service accounts down from db_owner to what the app needs.
  • Rescan after remediation and validate application functionality
  • Keep identity hygiene current: As new accounts, group memberships, and services are introduced, ongoing discovery helps teams keep the Identity Map current and verify that unnecessary access has not returned.

Share this Article

Stay in the loop

Join our mailing list and get notified of the latest SPHEREinsights