2017 · Italy · S - Other service activities · Data theft via web platform intrusion; e-voting disruption
Analysed 2026-11-06 by Hermes
Open this case in the tool →Data theft via web platform intrusion; e-voting disruption
Expertise to Execution
Evidence coverage: 5/6 stages, weight 34/60
Decisive stage: entry (clearest evidence) · Alternative: Strategy to Execution
(1) The Rousseau platform was the technical core of the Five Star Movement's direct-democracy model - it registered members, held donor data and ran internal votes - so the expectation at the victim was that a system exercising citizens' political rights was secured to that criticality; after the breach leadership itself asserted the platform's security was adequate or merely the teething trouble of a young startup. (2) In practice the Garante's inspection (provvedimento n. 548, 21-12-2017) found the portal and platform built on a CMS the producer had declared end-of-life on 31 December 2013 and whose vulnerabilities it had already published, passwords stored with weak algorithms or in clear text on the older blog, each vote still linked to the voter's phone number, and no sysadmin logging; an intrusion that came to light only through the attacker's own Twitter publication (rogue0, from 3 August 2017) exfiltrated the anagrafic database of members and donors, and further weaknesses were exploited during the September leadership vote despite a white-hat's reported SQL-injection flaw and claimed fix weeks earlier. (3) The classification rests on the entry stage (clearest evidence, primary): the expectations that failed were technical remedies that demonstrably existed - end-of-life notices from the producer, a researcher's report with the resolution, a structured development methodology that the regulator itself identified as the way to remedy the flaws - but they were never realised in the architecture and daily operations of an in-house-maintained proprietary stack; the knowledge sat with specialists (material) and the chain broke in execution, not in a decision leadership deferred or a warning it overrode. Remediation and notification conduct (public downplaying contradicted by the 2013 end-of-life date; prescribed fixes completed only after two extensions and a fine) is a secondary Strategy to Leadership pattern in the response stage, but that stage is contributing only and weaker-sourced, so it does not displace the entry stage. The entry vector itself is not load-bearing as an attacker technique - what matters is that the victim kept reachable, for four years after end of support, a system everyone with the expertise knew was unfixable in place. RM17 (Expertise to Execution lens): the maintainers could not replace or isolate the legacy Movable Type 4.31-en stack - obsolete since 2013 per its producer and maintained by a tiny in-house team - leaving the platform exposed to the very vulnerabilities the vendor had published (causal role: Enabling). RM29 (Expertise to Execution lens): secure development practice was known but not applied - the Garante attributed the exploited weaknesses to the absence of a structured development methodology that let feature speed override risk prevention, so injection-type flaws in the organisation's own build let the database be downloaded (causal role: Enabling). RM26 (Expertise to Execution lens): the operators could not bring password storage and stored-data protection up to standard (weak cryptographic algorithms, clear-text storage on the legacy blog, votes tied to phone numbers), so the exfiltrated database was readable and its exposure politically sensitive (causal role: Amplifying). RM33 (Expertise to Execution lens): no logging of database access and operations existed (shared non-nominal credentials through a web interface, MySQL community edition unable to record them), so the intrusion could neither be detected nor reconstructed from the victim's own instrumentation (causal role: Obscuring).
RM17 — Obsolete / unsupported technologyRM29 — Secure-development / non-production exposureRM26 — Data-protection / cryptographic weaknessRM33 — Logging / forensic-readiness gapVictim as written in the case file is 'Rousseau' (the platform); the Garante identifies Associazione Rousseau as the data controller (2017: Beppe Grillo as titolare, Associazione Rousseau as responsabile; titularity later moved to Associazione Movimento 5 Stelle) and Casaleggio Associati as operator - the case file's description matches the victim. EuRepoC scores the theft as non-sensitive (1 point), but the Garante records political data (art. 9 categories) with votes linked to voters' phone numbers. The disruption on the row reflects the disrupted September 2017 leadership vote, reported as part of the same hack sequence. Actor left unattributed per the dataset; secondary reporting (Freedom House, Logic) names the publiciser as black-hat @r0gue_0 and reports that the party pressed legal action against the white-hat whistleblower Gubello - not treated as fact for scoring. Reuters states the September election was dogged by hacking attacks; this is reported attribution of the disruption, cited as secondary. Sector: a political party's membership/e-voting platform maps to SBI section S (other service activities, membership organisations), not O, as the victim is a party association and not a state body. No related_rows in the case file; no prior result for this victim in the output directory. Link check: reuters.com blocks automated access (HTTP 401); the page was opened and read during research. Link check: france24.com blocks automated access (HTTP 403); the page was opened and read during research. Security policy: no leak sites or dumped data were accessed; all sources are regulator decisions, established press and research reporting.