← Breach analysis by the Dual Vulnerabilities framework

French Ministry of Finance G20 data theft (Unattributed, 2010)

2010 · France · O - Public administration and defence · Cyber-espionage data theft via spear-phishing trojan

Analysed 2026-11-06 by Hermes

Open this case in the tool →

What happened

Cyber-espionage data theft via spear-phishing trojan

Dual vulnerability type

Expertise to Execution

Evidence coverage: 4/6 stages, weight 33/60

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

Why this classification

(1) The Ministry of Finance held the rotating G8/G20 presidency from November 2010, making the protection of the Treasury and presidency-preparation systems a recognised national concern - the 2008 Livre blanc had made cyber defence a national priority and ANSSI had been created in 2009 to raise the state's information systems security. (2) In practice the ministry's own technical estate did not implement the known baseline: users held administrator status on their workstations and installed booby-trapped PDF attachments from hijacked habitual correspondents; market anti-virus did not detect the spyware, specific surveillance devices that would have detected it were never deployed; the internet access gateway was misconfigured and the event logs it generated were never verified; there was no general network map, so it took weeks after the (external) alert to find all internet gateways; and data was exfiltrated, fractioned into normal traffic, undetected for about three months. Detection came only when foreign G20 correspondents reported the booby-trapped attachments in late December 2010, not from the ministry's monitoring. (3) The classification rests on the detection stage, where the Senate's formal investigation report (r11-681) documents the mechanism explicitly: the remedy existed and was knowable to technical specialists - deployable surveillance devices, log review, privilege tiering, asset mapping - and the chain broke in operational implementation; no source shows Bercy leadership having been warned and deferring or underfunding these controls, so the gap sits within the material column: expertise available but not executed (Expertise to Execution). The entry vector (spear-phishing) is documented but not load-bearing beyond the privilege condition; attribution to China is media-based speculation and not load-bearing. A secondary doctrinal pattern exists - the report's general finding that many ministries lacked a real security policy and that security was seen as a costly constraint - but for Bercy no specific doctrinal artefact or leadership decision is documented, so it remains uncoded and the runner-up. RM28 (Expertise to Execution lens): endpoint protection existed but was not effective where it mattered - anti-virus coverage could not detect the spyware without the specific surveillance that was never deployed, and application control was absent because users kept administrator rights and installed attachments freely (causal role: Enabling). RM34 (Expertise to Execution lens): the tools and telemetry were there but were insufficiently configured, staffed and maintained - the internet gateway was misconfigured, its logs were generated but never analysed, and no specific monitoring was in place, so a month-old intrusion only surfaced via external correspondents (causal role: Obscuring). RM15 (Expertise to Execution lens): the technical teams could not scope or protect the estate because exposure paths were unknown - no general network map existed and it took several weeks to identify all internet gateways while the trojan spread from workstation to workstation (causal role: Enabling). RM35 (Expertise to Execution lens): egress visibility was absent, so documents exfiltrated in fractions hidden inside normal traffic left the ministry undetected for about three months (causal role: Amplifying).

Evidence by stage

Risk markers

Sources

Notes

EuRepoC row; victim, country (France) and year (2010, start_date 01.12.2010) from the case file. Attribution to China is media-based speculation only (eurepoc: attribution_basis Media-based attribution, initiator_country China, never officially confirmed) and is not load-bearing. Account of the first alert differs slightly across sources: the Senate report says foreign G20 correspondents of the direction du Trésor reported the booby-trapped attachments on the nights of 30-31 December; WSJ (via IEEE Spectrum) says a ministry employee noticed a mail sent in their name that they had not sent. Both support the same detection gap: not internal monitoring. Disclosure was made by media (Paris Match, 07/03/2011) rather than by the ministry itself; the Senate report nevertheless treats the decision to make the attack public as a deliberate, welcome choice, so no disclosure risk marker is coded. No related_rows in the case file. Link check: france24.com blocks automated access (HTTP 403); the page was opened and read during research.