← Breach analysis by the Dual Vulnerabilities framework

Central European defence companies Cyber-espionage (suspected Lazarus, 2019)

2019 · O - Public administration and defence · Cyber espionage via LinkedIn spearphishing with business email compromise attempt

Analysed 2026-11-06 by Hermes

Open this case in the tool →

What happened

Cyber espionage via LinkedIn spearphishing with business email compromise attempt

Dual vulnerability type

Expertise to Execution

Evidence coverage: 3/6 stages, weight 19/60

Decisive stage: detection (clearest evidence) · Alternative: Strategy to Execution

Why this classification

(1) The expectation at two defence and aerospace companies handling technical and business-sensitive information was a material one: their environments should generate forensic logs (PowerShell logging), detect scheduled-task persistence, AD enumeration and password brute-forcing against administrator accounts, and monitor outbound traffic to cloud storage. (2) In practice the September-December 2019 intrusion ran entirely under the victims' radar: entry via fake LinkedIn recruiters, WMIC/scheduled-task persistence, AD enumeration followed by admin-account brute-forcing, WMI lateral movement, data archived and uploaded to Dropbox with a custom dbxcli client, and a final BEC attempt through a compromised mailbox and a lookalike domain - none of it noticed by the victims. The campaign was uncovered only when ESET discovered it at the end of 2019 and, on the victim side, only when the targeted customer contacted the correct email address instead of paying. PowerShell command logging was disabled by default, so even ESET's forensic reconstruction could not retrieve the executed commands or determine what data was taken. (3) This is a knowing-doing gap inside the material column, not a doctrinal one: no policy, strategy or leadership decision is documented anywhere in the case record, while every remedy the evidence points to (enable logging, alert on directory enumeration and admin brute-forcing, monitor egress to cloud storage) sits with technical specialists and was simply not realised in operations. The classification rests on the detection stage (clearest evidence, load-bearing); the entry vector itself is well documented but carries no evidenced victim-side condition (no source speaks to training, verification or endpoint-protection practice), and containment reads as the same execution gap only as inference (the escalation path is reported but never confirmed to have reached administrative control, so it is described here rather than coded). Recovery was not in play and the documented response conduct was sound. (4) RM33 (Expertise to Execution lens): experts could not investigate because the logs were missing - PowerShell logging was off by default on the compromised systems, so executed commands were unrecoverable and the scope of the theft stayed undetermined (causal role: Obscuring). RM34 (Expertise to Execution lens): the available detection tooling never surfaced a months-long intrusion that used detectable signals like scheduled-task persistence, AD enumeration and admin password brute-forcing; detection came from an outsider, not from the victims' monitoring (causal role: Obscuring). RM35 (Expertise to Execution lens): egress visibility for cloud-storage uploads was not in place, so archived company data could be pushed to Dropbox by the attackers' custom client without being detected or interrupted at a defence-sector company (causal role: Amplifying).

Evidence by stage

Risk markers

Sources

Notes

Victims unnamed: ESET cites client confidentiality and Reuters confirms two Central European defence/aerospace firms. Attribution is hint-based only - ESET found no strong evidence linking the operation to a known group, only Lazarus similarities (dataset initiator field lists the full Lazarus alias set as a dataset label, not incident-specific confirmation). Country left null: the dataset records only regions (Europe; Middle East) and the victims are unnamed - never guessed. Year 2019 from the dataset start_date (01.09.2019), matching the ESET report (attacks September-December 2019). All victim-side lifecycle evidence comes from one primary forensic report (ESET white paper, TLP:WHITE, produced with the cooperation of two victims); the BEC payment fraud was directed at the victim's customer, whose own verification procedure worked, so no business-process verification marker is coded. The dataset row records two receiver entries (Defence industry twice), consistent with the two-victim campaign. Link check: reuters.com blocks automated access (HTTP 401); the page was opened and read during research.