← Breach analysis by the Dual Vulnerabilities framework

Belarusian KGB personnel data theft (Cyber Partisans, 2023)

2023 · Belarus · O - Public administration and defence · Network intrusion with data theft and doxing

Analysed 2026-11-06 by Hermes

Open this case in the tool →

What happened

Network intrusion with data theft and doxing

Dual vulnerability type

Expertise to Execution

Evidence coverage: 4/6 stages, weight 14/60

Decisive stage: egress (clearest evidence) · Alternative: Expertise to Leadership

Why this classification

(1) The expectation at a national intelligence agency is that its own internet-facing systems are hardened against known weaknesses, that an intrusion into its own network is detected, and that the personnel database - the crown jewels that expose its officers to doxing and targeting - is monitored and protected; the agency had also been publicly alerted to exactly this threat when its chief admitted on state television in 2021 that there had been hacker attacks on its personal data. (2) In practice, the Belarusian Cyber Partisans announced on 26 April 2024 that they had exploited vulnerabilities in the KGB's web infrastructure (no zero-day involved per the EuRepoC record), claimed access dating back several years with the database taken in autumn 2023, and published admin lists, the website database and server logs, claiming personal files of over 8,600 current and former employees, later fed into a Telegram bot that identifies KGB staff from photos; the KGB website had been down for roughly two months, and the agency never commented, acknowledged the loss, or announced any detection. (3) The documented gaps - a known, remediable vulnerability left exploitable, a multi-year undetected presence in the network, and an undetected bulk departure of the personnel database - all sit inside the material column: the remedies (patching, monitoring, egress control) are standard operational practice, and the chain broke where operations failed to realise them. No source documents any doctrinal artefact (policy, programme, budget decision) of the victim that failed, so the gap is Expertise to Execution rather than a Strategy type; the classification rests on the egress stage, the most consistently reported across AP, RFE/RL, heise, The Record and Security Affairs, with entry load-bearing as well - the entry vector itself is only attacker-claimed and is treated as unverified. A secondary Expertise to Leadership pattern sits behind this: leadership publicly reframed the 2021 personnel-data attacks as foreign-intelligence work rather than remediating the weakness, and the comparable exposure recurred. That reading is the runner-up, but the recurrence stage is contributing rather than load-bearing, so the decisive evidence stays on the execution side of the material column. (4) RM18 (Expertise to Execution lens): the intrusion exploited known, remediable weaknesses in the KGB's web infrastructure that remained exploitable years after the group first gained network access, a remedy available to competent operations but never realised (causal role: Enabling). RM34 (Expertise to Execution lens): access claimed to date back several years, with the disclosure forced by the attacker's channel and no victim-side detection ever reported, shows monitoring and alert handling not operative against activity inside the agency's own network (causal role: Obscuring). RM35 (Expertise to Execution lens): files of over 8,600 employees left the network in bulk without being detected or interrupted, although egress visibility over the personnel database is a baseline expectation for exactly this data class at an intelligence service (causal role: Amplifying). RM21 (Expertise to Execution lens): the 2021 personnel-data leak, publicly acknowledged by the agency's chief, was not followed by implemented and validated remedies, and the comparable weakness recurred in 2023-2024 (causal role: Perpetuating).

Evidence by stage

Risk markers

Sources

Notes

Every mechanism claim in this case rests on the attacker's own announcement and its press reporting; the victim never confirmed any part of it. The Record states that the credibility and impact of the group's attacks are difficult to verify since victims rarely comment, but that Bellingcat and Belsat journalists have verified the authenticity of the group's data in the past; the 8,600-file figure comes from the group's coordinator speaking to AP. The EuRepoC row lists the group's Telegram posts (t.me/cpartisans_by/1403, /1408) as attribution sources; these were not opened, per the research rule to cite attacker claims only as reported by legitimate press. The KGB website has reportedly remained offline since April 2024 (Reform.news, 2026), which is why the recovery stage is recorded as not assessable rather than as a gap. Year 2023: the attacker claims the intrusion in autumn 2023 and the EuRepoC start_date is 01.01.2023; disclosure was 26.04.2024. The victim is the security apparatus of an authoritarian state and the actor an opposition hacktivist collective, but the victim is a real compromised organisation, not attacker-side infrastructure, so the DV is assessed on its lifecycle as usual; the politically motivated disclosure by the attacker means response/disclosure conduct could not be evaluated beyond the documented silence and the 2021 antecedent.