2022 · United States · O - Public administration and defence · Spear-phishing credential harvesting against government email accounts
Analysed 2026-11-06 by Hermes
Open this case in the tool →Spear-phishing credential harvesting against government email accounts
Expertise to Execution
Evidence coverage: 2/6 stages, weight 13/60
Decisive stage: entry (clearest evidence) · Alternative: Strategy to Execution
(1) DoE facilities run nuclear-energy, security and deterrence research - the indictment itself notes why their staff's accounts are prized - and against this very actor the CISA/FBI/NSA/NCSC joint advisory spells out the specialist remedies: phishing-resistant MFA, sender verification, email scanning and monitoring of mailbox rules. (2) In practice, per the federal indictment (N.D. Cal. 3:23-cr-00447-CRB), from 30 May 2022 employees at several DOE facilities walked through spoofed-colleague emails and fake PDF 'documents' into attacker-controlled fake login pages and gave up their government email passwords; the attackers logged in with the password alone, maintained access for months, and rewrote the inbox rule so IT-department alert mail went straight to the Deleted Items folder; Reuters showed the group standing up fake login pages for three national laboratories in the same Aug-Sep 2022 phase; nothing in the record shows DoE detecting the compromise itself, and it surfaced only through the December 2023 indictment. (3) The classification rests on the entry stage (detection points the same way): the unmet expectation was technical knowledge held by specialists - documented remedies for this exact campaign - that was not realised in the operations of the affected accounts, since a phished password alone bought persistent legitimate access and unnoticed mailbox-rule changes. The break is execution, not leadership: no source shows a DoE decision deferred, under-resourced or a warning discounted, so no leadership-side origin or break is evidenced, and Strategy to Execution (a named doctrinal policy failing) loses to Expertise to Execution on the absence of any documented policy artefact. The DV claim is the absent identity and monitoring controls in operation, not the attacker's use of phishing. RM22 (Expertise to Execution lens): government email accounts at DOE facilities stayed accessible on harvested passwords alone - the indictment logs the conspirators using those credentials to log in as legitimate users and maintain access for months, while agencies had documented MFA and credential/forwarding-rule monitoring as the available identity controls for this campaign (causal role: Enabling). RM24 (Expertise to Execution lens): the verification and awareness controls specialists prescribe did not hold in practice - across May-October 2022, staff at multiple DOE facilities repeatedly fell for spoofed colleagues, a 'missing attachment' bait they replied to, and PDF links to fake document-login pages (causal role: Enabling). RM34 (Expertise to Execution lens): persistent foreign logins and an inbox-rule change went unobserved for months - the concealment worked precisely because no alerting on anomalous logins or mailbox rules intervened - so detection did not identify the incident in time (causal role: Obscuring).
RM22 — Authentication / credential exposureRM24 — Human-factor / social-engineering susceptibilityRM34 — Monitoring / detection / alert-handling gapDescription matches the stated victim: the indictment (opened via pdftotext, case 3:23-cr-00447-CRB) confirms Callisto Group = FSB Centre 18, the May-October 2022 spear-phishing against DOE facility employees, credential harvesting via fake login pages behind PDF/attachment lures, and the DOJ press release of 5 Dec 2023 confirms the 2016-2022 targeting of DOE facility employees. Disclosure was law-enforcement-driven (~14 months after the campaign window); the DoE made no contemporaneous public statement. The Feb 2023 DoE statement (no evidence of information compromised) concerned the Aug-Sep 2022 attempted intrusions at Brookhaven/Argonne/Lawrence Livermore; public sources do not tie the successful May 2022 compromise of 'DOE Employee A' to those three labs, so the statement is not read as a false public claim (RM44/RM39 not coded). The detection stage's victim-side reading (DoE monitoring did not catch the compromise) is inference from the documented persistence and concealment; no source states a DoE detection failure. No related_rows in the case file; no earlier result for this victim found in the output directory. Group-wide pattern sources (Google TAG SPICA reporting via CyberScoop, Treasury release on 2FA-evasion tooling) used only as context, not as incident-specific claims. Link check: reuters.com blocks automated access (HTTP 401); the page was opened and read during research.