← Breach analysis by the Dual Vulnerabilities framework

Central Election Commission (CEC; Russia) et al. espionage intrusion (Red Stinger / Bad Magic, 2022)

2022 · Russia · O - Public administration and defence · Cyberespionage intrusion via spear-phishing and custom backdoor

Analysed 2026-11-06 by Hermes

Open this case in the tool →

What happened

Cyberespionage intrusion via spear-phishing and custom backdoor

Dual vulnerability type

Expertise to Execution

Evidence coverage: 2/6 stages, weight 12/60

Decisive stage: detection (clearest evidence) · Alternative: none defensible

Why this classification

(1) The expectation at election-related and civil-administration bodies whose staff ran the 2022 referendums in Donetsk, Lugansk and Crimea - and at the Russian Central Election Commission, whose advisor was among the targets - is a baseline one: endpoints protected by common antimalware and monitored well enough to notice a persistent backdoor on the machine of an election official. (2) In practice the opposite held: a decoy-document/LNK delivery chain ran end to end and installed the PowerMagic backdoor and the CommonMagic framework, which persisted via a daily scheduled task and used OneDrive/Dropbox folders as command and control; Kaspersky found the active infection only in October 2022 (campaign material since September 2021) and Malwarebytes dates the group's operations to December 2020, calling them undetected for at least three years - discovered by vendors, never by any victim, and never acknowledged by one. (3) That is a knowing-doing gap inside the material column, not a doctrinal one: the remedy was ordinary and available (Kaspersky reports the malware and techniques are not particularly sophisticated, and Malwarebytes documents an ordinary antimalware scan detecting and cleaning a variant of the same backdoor on one target), so the knowledge sat with technical specialists and simply was never realised in the victims' day-to-day operations; no source documents any victim-side policy, leadership decision, deferred programme or ignored warning, so no doctrinal or leadership reading is available, and the origin/break answers are explicitly inferred from these documented facts. The classification rests on the detection stage (longest-documented dwell, external discovery); the entry gap (endpoint chain unblocked) is corroborating; the initial vector itself is undisclosed (Kaspersky: the initial vector of compromise is unclear) and is not load-bearing. (4) RM34 (Expertise to Execution lens): available detection tooling was evidently not deployed or tuned at these organizations, since a persistent, non-novel backdoor ran unnoticed for months to years and only external vendor telemetry ended the dwell time (causal role: Obscuring). RM28 (Expertise to Execution lens): endpoint and malware protection was absent or ineffective across the targeted administrations' machines - the full MSI/VBS/PowerShell infection chain and its persistence installed unimpeded at several organizations, while on a comparable target an ordinary antimalware scan did detect and clean the same backdoor family, showing the protection shortfall was executional rather than an undetectable evasion (causal role: Enabling).

Evidence by stage

Risk markers

Sources

Notes

Multi-receiver EuRepoC row: the victim string carries six receivers (CEC Russia; Yasinovataya administration; DPR administration, port of Mariupol; three Not available) and matching receiver_country values Russia;Ukraine x5 - the Ukraine entries are occupied-part-of-Donetsk-region receivers; Country is set from the primary (first-listed) victim, CEC Russia. The six receivers line up with the five OP#5 targets Malwarebytes lists (Yasinovataya, Mariupol port, CEC advisor, transport-linked target, Vinnitsa library). All lifecycle evidence comes from security-vendor forensic reports (Kaspersky, Malwarebytes): no victim statement, regulator finding or court document exists, so the material-origin/execution-break reading of the gaps is inferred from the vendors' documented facts and confidence rests at 3. Impact as recorded by EuRepoC is minimal (weighted_intensity 1, hijacking without misuse, no disruption, no confirmed data breach at this incident), although the vendors document surveillance collection across the group's operations. Actor attribution is technical only (Kaspersky and Malwarebytes find no link to known groups; state sponsorship unestablished) - the unattributed actor does not lower the DV confidence, which is assessed victim-side. The initial access vector is undisclosed by both vendors and is not load-bearing. related_rows (228, 361, 3372, 3414) carry titles describing other incidents (CSTO defacement, Wagner site hack, GLONASS leak, IT Army DDoS) that do not share this row's victims; no earlier incident is documented for the receivers of this row.