Forensic Case Study: A Compromised Account Used as a Phishing Relay

Scope the recipients, not just the mailbox

A taken-over mailbox is serious by itself. Mail content can be read, rules can be planted, and payment fraud can follow. In a phishing relay incident there is often a second scope that gets overlooked: the people who received lures from that trusted address. This composite case study shows how responders typically think about both.

Composite case study: based on patterns observed across incident response engagements; details are blended and illustrative.

The situation

Relay incidents commonly start the same way. An attacker takes over a mailbox and uses it to send shared-file notifications, guest invitations, or document links to the people that account normally talks to. Recipients see a sender they know and trust, with a normal signature, so the lure is more likely to be opened.

Organizations can learn about it from outside. A recipient, such as a partner, reports a suspicious shared document. In other incidents the mail platform blocks outbound sending, or a spike of new guest accounts appears in the directory. The elements below are presented as common patterns across relay incidents, not as one sequence of events.

Signals that commonly cluster

The signals typically cluster around one account:

  • A burst of outbound mail or guest invitations in a short window.
  • Bulk deletion of sent items, and deleted-item recovery activity on the mailbox.
  • A rule that hides replies, bounces, or out-of-office messages, so the account owner sees nothing unusual.
  • Unusual external sharing in the file platform.
  • External recipients the account had not contacted before.

Audit records can help here, but they depend on logging being enabled and on retention. If the sent trail has been deleted and audit data is thin, the picture can be incomplete, which is one reason message trace matters.

How responders typically confirm scope

Responders typically establish a window from the first anomalous send to the last. Message trace for the sender typically gives the recipient list within its retention window, even when the sent folder has been emptied; older activity needs a different search, because trace retention is limited. Message trace covers mail sent from the mailbox. Guest invitations are typically not sent from the mailbox, so they do not appear in its trace; scope them and external shares from the directory audit logs and the file platform audit logs. Sharing activity and guest accounts created during the window are listed next. The result is a scope built around who received what, not just which account was touched.

What these incidents can cost

The cost in these incidents is commonly measured in time and trust. Teams may spend considerable effort rebuilding a sent trail that was deleted, and external contacts may be less willing to open a shared link for a while.

Containment

Containment starts as soon as the problem is understood, and it is not held for scoping. Disable and secure the account, revoke sessions, rotate credentials, and review registered MFA methods and devices, OAuth app consents, mailbox forwarding settings and admin role changes, to reduce the chance the intruder can regain access once the account is re-enabled. Remove the rule that hides replies as well. Evidence capture runs in parallel: record or export the rule details (conditions, actions, creation time) before removing the rule, and export message trace and preserve audit data while those steps happen.

What should wait is the clean-up of sent mail. A common first wrong move is purging the mailbox before the recipients are scoped. That removes the very trail needed to know who to warn. Collect the recipient list first, then clean up.

From there, responders typically:

  • Remove external sharing links and pending invitations.
  • Review guest accounts created during the window, since they can persist after the mailbox is cleaned.
  • Remove the lure where you can. Inside your own tenant, an administrator can search for the message and remove it. For external recipients, tell them and ask them to delete the message. Doing this quickly tends to help, so start as soon as the recipient list exists.
  • Notify affected recipients with plain guidance, then watch for follow-on compromise on their side.

Closing the incident without telling recipients can be a costly miss. Their clicks and sign-ins become a separate investigation, and they cannot start it if nobody tells them.

Lessons

Internal trust is an attack surface. A lure from a colleague is more likely to be opened than one from a stranger, so limit who can invite guests and share externally, and rate-limit outbound sending.

In response, secure the account at once and do not wait on scoping to do it. Preserve the sent trail before deleting it. Then treat the recipients as part of the scope, because they are often the easily overlooked next step.

Checklist for your next incident

  • Disable the account, revoke sessions, and rotate credentials immediately.
  • Review registered MFA methods and devices, OAuth app consents, forwarding settings and admin role changes.
  • Record the details of hidden-reply rules, then remove them, and review other mailbox rules.
  • Export message trace and audit data in parallel with containment.
  • Hold off deleting sent mail until the recipient list is captured.
  • List external shares, links, and guest accounts created in the window.
  • Remove sharing and invitations, then request removal of the lure from recipient mailboxes.
  • Tell affected recipients what happened and what to do.
  • Review who can invite guests and share externally.

Message us to talk through your phishing relay response plan.


Discover more from Core-Infosec

Subscribe to get the latest posts sent to your email.

Picture of Core Watch

Core Watch

Practical security guidance from Core-Infosec.

Read More