2013 · Japan · J - Information and communication · Data theft via administrative system intrusion
Analysed 2026-11-06 by Hermes
Open this case in the tool →Data theft via administrative system intrusion
Expertise to Execution
Evidence coverage: 1/6 stages, weight 6/60
Decisive stage: egress (clearest evidence) · Alternative: Strategy to Execution
(1) Yahoo Japan operates Japan's largest web portal, with roughly 200 million members and over 24 million active IDs held in its back-end systems; the expectation that such an operator can establish whether a customer-data file has left its network - file-level egress monitoring and forensic scoping - is a settled technical capability of its own specialist staff (material origin). (2) In practice, when an unauthorized attempt against its administrative system was detected late on 16 May 2013, the company cut net access and investigated, but could only conclude from the volume of traffic between the back-end admin system and the internet that a file of 22 million user IDs had possibly been taken; it publicly stated it did not know for sure the file had been taken, never disclosed how the access happened, and no later public investigation outcome is on record. (3) This is a knowing-doing gap inside the material column: the remedy was known technical practice, not a strategic commitment, and no source shows leadership deciding, deferring or under-resourcing it - the capability was simply not realised in operations, so the chain broke at execution, giving Expertise to Execution. The egress gap is load-bearing: the incident's defining impact - a 22-million-record customer file possibly copied out - was never detected or interrupted at transfer level, and remained an unconfirmed inference. The classification rests on the egress stage; the entry vector is unknown and not load-bearing, and detection (internal, same evening), containment (immediate net-access cut; the file held no passwords) and response (prompt, self-initiated disclosure with an ID-check tool) are documented as aligned. A secondary pattern shows at the same victim in April 2013, when malware on company servers had collected 1.27 million account records before being stopped, but its mechanism is not comparably documented, so it is not coded. RM33 (Expertise to Execution lens): Yahoo Japan's investigators could not scope or reconstruct the incident at file level - they could not determine what data was taken, only measure aggregate traffic - because the logging and forensic tooling to confirm the file's movement did not exist in operations (causal role: Obscuring). RM35 (Expertise to Execution lens): bulk outbound traffic between the back-end admin system and external terminals was noticed only after the fact through its volume; egress visibility and prevention capable of detecting or interrupting the copy of a 22-million-record customer file were not implemented (causal role: Amplifying).
RM33 — Logging / forensic-readiness gapRM35 — Egress / exfiltration monitoring gapThe dataset source_url (Reuters/Kyodo, 17 May 2013) carries a mangled trailing ' /' in the case file and returns HTTP 403 to automated access, so the analysis rests on the articles actually opened and listed in Manual source-article. The intrusion vector was never disclosed ('The company did not say exactly how the unauthorized access took place or how long the attack may have taken' - Dark Reading) and no public investigation conclusion on whether the file left was found in the reporting reviewed. Context: an earlier incident at the same victim in April 2013 - malware on company servers extracted user data for 1.27 million users but was stopped before exfiltrating it; SecurityWeek reported the malware was allegedly planted via a backend application vulnerable to Remote File Inclusion, a detail the outlet itself could not confirm; the 'Secret ID' login feature was added in the upgrade after that incident. RM21 is not coded for the May case because neither incident's mechanism is documented well enough to establish a comparable recurring weakness. The stolen data was user IDs, which Yahoo Japan emphasized are already public information, and no passwords were in the file; the impact claim (22 million IDs possibly leaked) rests on the company's own traffic-volume inference and was never confirmed.