2019 · Bahrain · B - Mining and quarrying · Destructive cyberattack with data-wiping malware
Analysed 2026-11-06 by Hermes
Open this case in the tool →Destructive cyberattack with data-wiping malware
Expertise to Execution
Evidence coverage: 3/6 stages, weight 21/60
Decisive stage: containment (clearest evidence) · Alternative: none defensible
(1) Bapco is Bahrain's state-owned national oil company in the Gulf energy sector that Iranian wiper campaigns had been hitting for years (Shamoon at Saudi Aramco 2012, Shamoon v2/v3, ZeroCleare 2019); the standing expectation, spelled out in the Saudi NCA's own tactical recommendations to the sector, is that internet-facing VPN appliances stay patched, service accounts stay out of Domain Admin groups, and management tooling reach is constrained. (2) What happened: the NCSC assessed with moderate confidence that attackers broke in through a VPN appliance remote-execution vulnerability disclosed in July 2019 and held a foothold from summer 2019; inside, they obtained domain admin and service accounts, copied PsExec and Dustman onto the antivirus management console server - connected to every machine by the nature of its functionality - distributed the wiper across all connected systems, manually deleted the storage server's volumes, and erased VPN logs; Bapco noticed only the next morning when antivirus on sleeping workstations blocked the payload. (3) The classification rests on the containment stage: the remedies for every documented gap are known technical practice (material origin), and nothing in the sources shows a leadership decision deferred or under-resourced them - the gap is that known practice was never realised in the operational architecture, with flat privileges and a single management console as the distribution path for the whole fleet (execution break). The entry gap (an unpatched, publicly disclosed VPN bug) is the same type and load-bearing; the roughly five-month undetected dwell is a contributing, inference-grade gap; the entry vector itself is moderate-confidence and is not where the DV reading rests. No data was stolen and operations continued, so egress and recovery are not in play as gaps. Secondary pattern worth naming: disclosure sat entirely with the Saudi NCA and press, since Bapco never publicly acknowledged the incident - but no source evidences a doctrinal expectation or leadership decision, so no second DV-Type is defensible. RM18 (Expertise to Execution lens): the VPN remote-execution vulnerability was known and remediable - disclosed in July 2019 with vendor fixes - yet the internet-facing appliance stayed exploitable into the summer 2019 compromise, so the patch remedy existed but was not executed (causal role: Enabling). RM41 (Expertise to Execution lens): privilege tiering and directory hygiene were not realised - the attackers obtained domain admin and service accounts through configuration weakness, not a zero-day, giving them domain-wide control (causal role: Amplifying). RM23 (Expertise to Execution lens): Bapco's own antivirus management console, whose reach spanned all machines and whose service account was not isolated, was turned against it as the wiper's distribution mechanism, so the blast radius equalled the tool's reach (causal role: Amplifying). RM34 (Expertise to Execution lens): the tooling available to the security function did not surface an interactive intrusion for roughly five months, and detection happened only via endpoint antivirus the morning after detonation, an obscuring detection-configuration gap rather than missing logs (causal role: Obscuring).
RM18 — Vulnerability / patch latencyRM41 — Privilege-escalation / directory-hygiene pathRM23 — Management-tooling weaponisationRM34 — Monitoring / detection / alert-handling gapCase-file description matches the incident researched (Bapco, data-wiping malware, Iran attribution). Dataset dates: start 01.01.2019 / end 29.12.2019; the destructive payload detonated on 29 December 2019 and the compromise itself dates from summer 2019 (the EuRepoC analysis, the target of the case file's bit.ly link, gives the window 25 July to 29 December 2019). The primary source, the Saudi NCA/NCSC technical report Destructive Attack DUSTMAN (TLP:WHITE), was read through the public Scribd copy carried as a case-file source_url (uploaded by the ZDNet journalist); parts of that page render as images, but the readable text confirms the lifecycle (VPN exploit, credential harvesting to domain admin/service accounts, months-old signs of compromise, VPN-log deletion) and ZDNet quotes its key passages verbatim. Attribution is third-party and unverified (dataset: anonymous statements in media; ZDNet: security experts could not link the attack to a specific Iranian group) - an unattributed actor does not lower the DV confidence. Entry vector assessed by NCSC with moderate confidence; ZDNet's sources differed on the appliance vendor (Fortinet vs Pulse Secure); this does not affect the DV reading. Disclosure was third-party: Bapco made no public statement and did not respond to press requests; RM39 was considered and not coded because no source documents an unreasonable, inaccurate or concealed disclosure by the victim. The Jerusalem Post link in source_url covers Albania 2024 and general ZEROCLEARE background; it was opened and is not used as evidence for this case. Sector: national oil company (petroleum concession) - SBI section B. Unattended run: nothing encountered required consent.