2023 · United States · O - Public administration and defence · Ransomware attack with pre-encryption data theft
Analysed 2026-11-06 by Hermes
Open this case in the tool →Ransomware attack with pre-encryption data theft
Expertise to Execution
Evidence coverage: 6/6 stages, weight 49/60
Decisive stage: entry (clearest evidence) · Alternative: Strategy to Execution
(1) The City of Dallas, operating critical infrastructure for the ninth-largest US city (police CAD, courts, utilities, health plans), had a funded and doctrinally complete security programme: security spending rose from USD 3.4 million in 2019 to USD 7.8 million in 2023, a Security Operations Center, EDR and AI-aided detection were deployed, a cloud identity platform extended MFA to user logins across 200+ applications, backups had been restructured after 2021 events, and the Incident Response Plan was exercised - the expectation being that every authenticated account is protected, intrusion activity is surfaced, bulk egress is noticed and the public is told accurately what was taken. (2) In practice, Royal used a stolen domain service account to log in on 7 April 2023; the account sat outside the MFA-covered identity platform, and the attacker spent 27 days harvesting credentials, traversing the estate with legitimate third-party remote management tools and beacons, and exfiltrating an estimated 1.169 TB through the one compromised server without detection; monitoring alerted the SOC only as encryption started on May 3 (ransom notes on 996 hosts, 230 servers impaired, police CAD disrupted), containment took about a day, restoration from backups ran May 9 to June 13 for about USD 8.5 million, and leadership continued telling the public there was no evidence data had been compromised until after its own June 14 forensic determination, announcing only on August 3. (3) The classification rests on the entry stage (clearest evidence, documented primary in the city's own after-action report): the unmet expectation was material - protecting and monitoring service-account credentials is established technical expertise, which the city demonstrably possessed for user logins - and the chain broke at execution, because the stolen service-account credentials still logged in and reached a server; no source shows leadership deferring or refusing that control. Containment, detection and egress show the same material-to-execution pattern (expertise available but not realised in architecture and operations: unhardened assets the AAR itself names, 27 days of undetected beaconing, undetected bulk egress). The response/disclosure gap is a doctrinal-to-leadership pattern - public statements contradicted by the city's own investigation - and is named as the secondary Strategy to Leadership reading; it does not decide the case because it rests on contributing-stage evidence while the load-bearing pre-breach entry gap is primary-documented (a pre-breach condition would in any case prevail over a post-breach stage). The specific route by which the service-account credentials were stolen is not disclosed; Royal's group-level callback-phishing pattern is pattern evidence only and is not load-bearing here. (4) RM22 (Expertise to Execution lens): the city's identity experts ran MFA and centralised authentication for 200+ applications, but the legacy domain service account stayed outside that coverage, so the authentication control in place could not protect the exact account Royal logged in with (causal role: Enabling). RM41 (Expertise to Execution lens): after that login, privileges and directory hygiene did not hold - Royal harvested users, accounts and devices and moved through the estate using legitimate remote-management utilities while assets were not consistently hardened - expanding the blast radius to an apparently prioritized server list (causal role: Amplifying). RM34 (Expertise to Execution lens): the deployed monitoring stack was not configured or tuned to surface a month of reconnaissance, beaconing and harvesting, alerting only at encryption - a 27-day time-to-detect (causal role: Obscuring). RM35 (Expertise to Execution lens): egress from the single initial server was unmonitored, so about 1.169 TB left the network while the telemetry to see it existed only in post-incident logs (causal role: Amplifying).
RM22 — Authentication / credential exposureRM41 — Privilege-escalation / directory-hygiene pathRM34 — Monitoring / detection / alert-handling gapRM35 — Egress / exfiltration monitoring gapThe case file's figure 1,169 TB misstates the after-action report, which gives an estimated 1.169 TB of exfiltrated data (decimal point, not thousands). The case file dates the AAR release to 21 September; the document itself is dated 6 September 2023 (final revision 20 September) and the city's Technology Accountability Report records publication on 20 September. The AAR is the victim's own report and frames the 27-day dwell time favourably against IBM industry averages; its forensic findings were nevertheless produced with external cybersecurity experts. How the service-account credentials were stolen is not disclosed; Royal's documented callback-phishing TTP (BleepingComputer) is group-level pattern evidence, not incident-confirmed, so no social-engineering marker is coded. The 30,253 notified individuals were mostly current, retired and prior city personnel and dependents via the city's self-insured health plans (26,212 Texas residents on the state-ag filing); the HHS Office for Civil Rights opened an investigation in August 2023 - no finding was available at analysis time. The disruption of the police CAD system is documented in the sources opened; the case file's link to the Allen, Texas mass shooting response could not be confirmed in those sources and is not relied on. All cited URLs were opened during this task.