← Breach analysis by the Dual Vulnerabilities framework

NISC Japan e-mail espionage breach (China-linked actor, 2022)

2022 · Japan · O - Public administration and defence · State-linked espionage intrusion with e-mail data theft

Analysed 2026-11-06 by Hermes

Open this case in the tool →

What happened

State-linked espionage intrusion with e-mail data theft

Dual vulnerability type

Strategy to Execution

Evidence coverage: 3/6 stages, weight 15/60

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

Why this classification

(1) NISC is Japan's national cyber command centre: it drives the national cybersecurity strategy, monitors cyberattacks against government organisations, and in March 2023 co-issued the national guidelines on sharing and publishing cyberattack damage information. (2) In practice an intrusion into NISC's own e-mail-related system ran from early October 2022 to mid-June 2023 (about eight months): traces of illicit communication were only found on 13 June 2023, the cause was a device vulnerability the manufacturer had not yet identified, mail data of about 5,000 correspondents may have leaked, and the public announcement of 4 August 2023 followed the Financial Times report while withholding the device and vulnerability details, for which JPCERT/CC pointedly referred back to the sharing guidelines. (3) The classification rests on the detection stage: the expectation to detect such an intrusion originated doctrinally - in NISC's own strategic mandate to monitor government cyber security - but it broke at execution, since the operational monitoring of its own environment did not realise that mandate for eight months and monitoring was strengthened only afterwards; nothing in the sources shows leadership deferring or under-resourcing it. Entry is not load-bearing (the exploited defect was unknown to the manufacturer at the time of intrusion, so no victim-side patch gap is evidenced), and egress monitoring failing to interrupt eight months of mail exfiltration is a secondary, partly inferred Expertise to Execution pattern behind the same failure. (4) RM34 (Strategy to Execution lens): NISC's monitoring and alert-handling requirements - the detection capability it operates for all of government - were not operated as intended in its own environment, leaving an eight-month dwell time on its own mail system (causal role: Obscuring). RM35 (Strategy to Execution lens): egress controls and data-transfer restrictions on the mail gateway were not translated into an effective operational barrier, so a few thousand correspondents' mail left unnoticed and only post-incident reconstruction established the leak window (causal role: Amplifying). RM39 (Strategy to Execution lens): the notification and sharing workflow NISC itself co-published in March 2023 was not operationalised for its own incident - disclosure came after media reporting, technical details were withheld and the national CERT was left without a technical report (causal role: Constraining).

Evidence by stage

Risk markers

Sources

Notes

The case file's description matches the victim. Attribution to China rests on unnamed officials quoted by the FT (paywalled, so not cited directly; secondary reporting opened here reproduces the quotes); the Japanese government has officially refrained from attributing, so the Actor is recorded as China-linked. The dataset's zero_days=No conflicts with NISC's own advisory, which attributes the cause to a device vulnerability the manufacturer had not confirmed (Japanese media: a newly discovered defect); no vendor, CVE or threat actor was ever officially named for this device - press reporting on Barracuda e-mail gateway abuse is not relied on. The original nisc.go.jp/news/20230804.html URL now redirects to the redesigned cyber.go.jp home page; the same advisory is cited at its current cyber.go.jp location. Disclosure was media-triggered (FT, then NISC's statement of 4 August 2023); EuRepoC records source_disclosure as media. The external specialised investigation body was never named and JPCERT/CC publicly stated it did not conduct it. Egress-stage origin/break rest on the coder's inference (no source addresses egress monitoring specifically); detection is the only load-bearing stage.