← Breach analysis by the Dual Vulnerabilities framework

Baltimore Public Schools ransomware data theft (Cloak, 2025)

2025 · United States · P - Education · Ransomware attack with data theft

Analysed 2026-11-07 by Hermes

Open this case in the tool →

What happened

Ransomware attack with data theft

Dual vulnerability type

Expertise to Execution

Evidence coverage: 2/6 stages, weight 10/60

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

Why this classification

(1) Baltimore City Public Schools holds the personal data of a school system serving roughly 77,000 students, and its own breach notice states it is committed to safeguarding the privacy and security of all personal information stored on its servers; under Maryland's notification law it owed prompt notice to affected individuals. On the material axis, endpoint and egress monitoring are standard, available remedies, and the district itself treated them as such the moment the incident ended. (2) Operationally, on February 13, 2025 attackers (linked by WBAL-TV reporting to the Cloak ransomware group, which left a ransom note; the district never officially attributed the attack) reached its IT systems and took documents on roughly 31,125 people - employee records going back to 2010 including I-9 and background-check files, plus records for about 1,000-1,150 students - without the transfer being detected or interrupted; some of the data surfaced on the dark web before any public word. Detection and containment were not the weak points: the district noticed the incident at once, kept control of its systems, paid no ransom and kept schools running. The entry vector was never disclosed and is not load-bearing; CrowdStrike's forensic findings were not published. (3) The decisive gap sits at egress: the remedy (endpoint/egress monitoring) was a known, available technical control that the district deployed network-wide immediately after the incident - installing EDR software and resetting all passwords - yet it had not been realized in operations beforehand, and nothing in the record shows a leadership decision to defer or underfund it; the chain therefore broke between technical expertise and execution, which is Expertise to Execution. The 10-week delay to public notification, with the affected count revised upward on notification day, is a genuine doctrinal-column gap (Strategy to Leadership) recorded at the response stage, but it worsened the aftermath without making the breach possible, so the egress stage is the only load-bearing one and is decisive. Conditions no marker covers are noted here rather than coded: no regulator or formal investigation had reported findings on the district's controls at the analysis date, so no 'finding of no failings' applies. (4) RM35 (Expertise to Execution lens): the district could not detect or interrupt the bulk departure of files on more than 31,000 people, and the EDR software it installed right afterwards shows the egress and endpoint visibility experts needed was simply not implemented in the architecture before the theft (causal role: Amplifying). RM39 (Expertise to Execution lens): the scope could not be established before disclosure - the notification figure rose from 25,000 to 31,125 the same evening letters were mailed and the public word came about 10 weeks after the intrusion - because the forensic scoping had to be completed before anything could be said, constraining how promptly and accurately victims could act (causal role: Constraining).

Evidence by stage

Risk markers

Sources

Notes

Case file is an informationisbeautiful row with only a one-sentence description; it names the victim 'Baltimore Public Schools' - the official name is Baltimore City Public Schools (City district, Maryland). The description (tens of thousands of employees and students exposed) matches the February 13, 2025 incident. Not to be confused with the November 2020 ransomware shutdown of Baltimore County Public Schools, a different district (the Inspector General found that county district partially responsible), nor the 2019 RobbinHood attack on Baltimore City government - the 2025 incident is a separate one and no earlier comparable weakness at City Schools was documented, so no learning-failure marker is coded. Actor attribution to Cloak comes from WBAL-TV investigative reporting (ransom note); the district did not officially attribute, so the actor is recorded as reported, and this does not affect the victim-side DV. The initial access vector and CrowdStrike forensic findings were never disclosed; the entry stage is recorded as not documented, not as a gap. That the district added EDR only after the incident is the coder's reading of the district's own post-incident statement, so the egress stage and the classification claims are marked inference. Some stolen data was reported posted on the dark web; per policy no leak site was visited and the claim is cited only as reported by legitimate press. As of the analysis date no regulator or formal investigation had published findings on the district's controls (the Maryland Inspector General for Education had received no complaints), so the 'regulator finding of no failings' rule does not apply. The task brief carried no extra details beyond the case file; nothing was found on law-enforcement action beyond the ongoing FBI/DHS collaboration. No consent-requiring actions were attempted in this unattended run. Link check: wbaltv.com blocks automated access (HTTP 451); the page was opened and read during research. Link check: baltimoresun.com blocks automated access (HTTP 403); the page was opened and read during research. Link check: capitalgazette.com blocks automated access (HTTP 403); the page was opened and read during research.