2018 · India · K - Financial and insurance activities · ATM skimming and cyber-espionage via banking malware
Analysed 2026-11-06 by Hermes
Open this case in the tool →ATM skimming and cyber-espionage via banking malware
Expertise to Execution
Evidence coverage: 3/6 stages, weight 17/60
Decisive stage: detection (clearest evidence) · Alternative: Strategy to Execution
(1) The doctrinal expectation at the victim side was explicit: Indian banks ran their ATM estates under RBI instructions, including the circular of 21.6.2018 requiring immediate action on ATM control measures with time-bound software upgrade, while the specialist remedies - patching, traffic monitoring, antivirus, tightened network and password policy - were known and available in the sector. (2) In practice, from at least August 2018 to September 2019 the Lazarus group's ATMDtrack sat on Indian banks' ATMs reading and storing inserted card data and shipping it to a hardcoded C2 domain hosted in India, and the related Dtrack RAT collected host data and exfiltrated it, yet the campaign was surfaced by Kaspersky's telemetry and attribution work rather than by any victim's monitoring; Kaspersky had not determined the initial infection route at disclosure and reported that targeted entities were failing to track traffic across the organisation. Sector evidence shows why the estate stayed reachable: 74% of public sector bank ATMs ran outdated or unsupported software, bankers said they were unlikely to meet at least some of the deadlines prescribed by the regulator, and after the campaign RBI found banks relying on shared third-party ATM switch service providers whose contracts contained no baseline cyber controls. (3) The classification rests on the detection stage, the only load-bearing and best-documented stage: the expectation that was not acted on was material (remedies held by technical specialists - traffic monitoring, antivirus, policy hardening), and the chain broke in execution, since the controls were simply not realised in operations and no source documents a leadership decision to defer them; months of undetected skimming and espionage followed. That is a knowing-doing gap inside the material column (Expertise to Execution), not generic weak security. The entry vector is unknown and not load-bearing. Secondary patterns: the entry-stage condition (unsupported ATM software against a known regulatory upgrade obligation that managers acknowledged they would miss - doctrinal origin, leadership break) is a sector-level Strategy to Leadership pattern, and the response-stage assurance finding (vendor-assurance doctrine not contractualised) is a Strategy to Execution pattern; both are documented only at ecosystem level and are recorded as contributing, not decisive. RM34 (Expertise to Execution lens): available detection tooling was insufficiently configured, staffed or maintained to catch ATM-borne skimming and RAT traffic - the affected entities lacked traffic monitoring and the campaign ran for over a year before an external vendor surfaced it (causal role: Obscuring). RM17 (Expertise to Execution lens): the known remedy of replacing or isolating legacy ATM software was not executed, with 74% of public sector bank ATMs on outdated or unsupported software, the classic case of experts unable to retire legacy under operational dependency and capacity (causal role: Enabling). RM03 (Expertise to Execution lens): RBI's June 2018 time-bound ATM upgrade obligation was known before the campaign but not met - compliance requirements were understood yet not implemented (causal role: Predisposing). RM13 (Expertise to Execution lens): banks depended on shared third-party ATM switch application service providers that technical teams could not verify, because contracts imposed no baseline cyber controls until RBI mandated them after the campaign (causal role: Enabling).
RM34 — Monitoring / detection / alert-handling gapRM17 — Obsolete / unsupported technologyRM03 — Regulatory / stakeholder obligation gapRM13 — Third-party / supply-chain trust boundaryEuRepoC row carries no receiver_name; victim named descriptively from receiver_category Finance + receiver_country India. 'Lazarus returns against ATMs' is the EuRepoC incident name, not a victim. The description spells the malware 'ATMDTrack'; Kaspersky's forensic report names it ATMDtrack (subset of the Dtrack family). No individual bank is publicly named for this campaign - disclosure came from an IT-security company (EuRepoC source_disclosure), so victim-side gaps rest on Kaspersky's forensic report (its victim-condition statement is hedged: the security issues are inferred from attacker success) plus sector-level primary documents (Ministry of Finance reply to Parliament, RBI circulars) tied to the row's victim population, not incident-confirmed for any single bank. Year 2018 taken from eurepoc_extra.start_date (ATMDtrack discovery at Indian banks, late summer 2018); Dtrack activity continued to September 2019. Actor attribution is by technical report (code reuse from the 2013 DarkSeoul campaign); EuRepoC lists initiator as Unknown with state-affiliated category. The TNW article in source_url no longer opens (HTTP error) and was replaced by the Kaspersky report and the ISMG article; the CISA AA20-239A advisory (row's attribution source) covers the related BeagleBoyz/FASTCash 2.0 bank campaign and is cited as actor context, not as Dtrack-specific evidence. No related_rows; results for other Lazarus victims in the output directory (rows 13, 213, 2429, 2483) concern different incidents. Unattended run: nothing required consent.