← Breach analysis by the Dual Vulnerabilities framework

Université Paris-Saclay ransomware (RansomHouse, 2024)

2024 · France · P - Education · Ransomware attack with encryption and data theft

Analysed 2026-11-06 by Hermes

Open this case in the tool →

What happened

Ransomware attack with encryption and data theft

Dual vulnerability type

Expertise to Execution

Evidence coverage: 2/6 stages, weight 16/60

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

Why this classification

(1) Université Paris-Saclay, a top-ranked French public university processing large volumes of student and applicant personal data, is expected under its own public-service continuity obligations and GDPR/CNIL duties to know what data it holds, contain an intrusion and be able to establish what left its network; ANSSI's 2024 annual report explicitly cites this incident for the importance of continuity and recovery plans. (2) In practice the 11 August 2024 WhiteRabbit ransomware encrypted all internal servers of the central services and its faculties, IUT, Polytech and OSU, knocking out centralised authentication, e-mail, intranet, shared spaces and business applications during the enrolment period and for months after, while the DSI - even assisted by ANSSI - could neither confirm nor deny a data breach for about two months and could not know the extent or nature of any data taken; the exposure surfaced only when the attacker published files on 9 October. (3) The classification rests on the egress stage, where the victim's own disclosure documents the gap: the knowledge and capability that was not realised was material - forensic scoping, data inventory and exfiltration reconstruction sit with technical specialists - and the chain broke at execution, because the specialists (DSI with ANSSI) could not realise it in operations; nothing in the record shows leadership deferring, underfunding or ignoring an expert warning, which would have pointed to Expertise to Leadership, and no documented policy or leadership decision sits behind the containment state, which keeps the Strategy-to-Execution reading inferential. The entry vector was never disclosed and is not load-bearing. The secondary pattern visible is a containment/segmentation blast-radius problem in the shared central estate. (4) RM33 (Expertise to Execution lens): the specialist teams could not investigate because the forensic basis to scope the theft was not there - the university itself stated for two months it could not confirm or deny a leak or know its extent or nature, and the taken-data question was answered only by the attacker's own publication (causal role: Obscuring). RM30 (Expertise to Execution lens): compartmentalisation was a known measure that demonstrably worked inside this estate (it preserved the hosted anti-doping laboratory), yet experts did not or could not realise it across the central architecture, so one intrusion reached every internal server of the central services and components and the outage ran weeks to months (causal role: Amplifying). RM04 (Expertise to Execution lens): many critical services hung on single central dependencies - a centralised authentication service whose reconstitution was a top recovery priority, and shared platforms such as the Cirrus shared space and mutualised infrastructure whose loss also disrupted member schools and partner research organisations - so one compromise forced wide outages (causal role: Amplifying).

Evidence by stage

Risk markers

Sources

Notes

Attribution: RansomHouse self-attributed the attack and claimed 1 TB of stolen data on its leak site; that volume is an unverified attacker claim, reported here only via press (LeMagIT) - the university confirmed only 153 leaked files from 44 Master's applications. ANSSI names the ransomware (WhiteRabbit, used by RansomHouse) but not the group in the passage read. The victim's FAQ page (universite-paris-saclay.fr/en/piratage) now returns 404 and is cited via a Wayback snapshot opened during this task (page last updated 16 September 2024). Dataset data-quality: the row's eurepoc_extra intelligence_impact says no data breach/exfiltration, which contradicts the confirmed October leak - the field predates the attacker's publication. The entry vector was never disclosed (ANSSI was asked to identify the enabling flaws; findings not public), so entry is recorded as a gap-less not-documented stage. No related_rows in the case file. Unattended run: nothing requiring consent was encountered; only public sources were read.