2021 · Russia · O - Public administration and defence · State-sponsored cyber espionage intrusion via spear-phishing (backdoor implant)
Analysed 2026-11-06 by Hermes
Open this case in the tool →State-sponsored cyber espionage intrusion via spear-phishing (backdoor implant)
Expertise to Execution
Evidence coverage: 2/6 stages, weight 12/60
Decisive stage: detection (clearest evidence) · Alternative: Strategy to Execution
(1) The institutes are the core of Russia's radio-electronics defense R&D (electronic warfare, radar, airborne satellite communications) inside the state-owned Rostec conglomerate; the expectation that applies to them sits with their own security specialists: e-mail and endpoint controls that stop macro-based spear-phishing droppers, and monitoring that finds an implant writing scheduled tasks, Run keys and beaconing to attacker domains. No public doctrinal artefact (policy, standard, programme) of the institutes exists in the record. (2) In practice, CPR's forensic report documents the chain executing end to end on the targets' machines: spoofed Ministry-of-Health lure, remote macro template, DLL side-loading via a signed cmdl32.exe, SPINNER beaconing to C&C infrastructure - with an earlier dropper wave active since at least June 2021 and the campaign still observed in April 2022. The intrusion was exposed by Check Point in May 2022; Rostec and Russian officials never acknowledged it, and no exfiltration was confirmed (EuRepoC codes hijacking without misuse). (3) The classification rests on the detection stage, not on the entry technique: what is documented is not that a novel, heavily obfuscated backdoor defeated signature-based defences (SPINNER was previously undescribed, so that alone is not a victim-side gap), but that an implant with plain, OS-visible persistence artefacts beaconed from workstations of a top-tier defense conglomerate for roughly ten months without its own monitoring noticing - specialist-knowable detection capability that was never realised in operations. The origin of the unmet expectation is material (technical detection and verification capability) and the break is in execution (operations did not deliver it); no source documents a leadership decision to defer, under-resource or accept the gap, which is why the leadership-side types are not chosen, and no doctrinal artefact exists to make it a Strategy-to-Execution diagonal. (4) RM24 (Expertise to Execution lens): the social-engineering risk that CPR and NYT describe - sanctions-themed lures posing as an official ministry document, delivered to scientists and engineers at defense institutes - was not met with effective verification procedures or training on the ground, since the lure worked repeatedly across waves from 2021 into 2022, showing the training/verification control specialists know how to build was not realised (causal role: Enabling). RM34 (Expertise to Execution lens): whatever monitoring and endpoint tooling the institutes had was not configured, tuned or staffed to detect a persistent backdoor's artefacts and periodic beacons for roughly ten months - the incident surfaced through an external vendor's telemetry, not their own alert handling (causal role: Obscuring).
RM24 — Human-factor / social-engineering susceptibilityRM34 — Monitoring / detection / alert-handling gapEuRepoC row; victim written as 'Rostec;Rostec' (duplicate of the same name); the targeted entities are research institutes within a radio-electronics holding inside Rostec. The dataset codes the incident as hijacking without misuse with no confirmed data breach or disruption, so egress and recovery stages are not in play and the response stage is undocumented: Rostec/Russia never publicly acknowledged the intrusion (CCDCOE toolkit citing later NYT reporting), which prevents coding any response-stage marker rather than evidencing alignment. Start year taken as 2021 from eurepoc_extra.start_date (01.07.2021); CPR's own dating of the earlier dropper wave is June 2021, and the March 23, 2022 wave is the one named in the row description. The victim-side origin/break answers for entry and detection rest on reading CPR's forensic chain plus the documented ten-month dwell; the victim itself publishes nothing, so no primary organisational source exists and the stage source types are inference. Link check: nytimes.com blocked direct automated fetch during research; the same article's full English text was opened and read on the official NYT bilingual mirror cn.nytimes.com, cited instead. No regulator, court or victim statement exists for this incident, so confidence cannot exceed the inference level of the classification claims. Unattended run: nothing requiring consent was encountered.