← Breach analysis by the Dual Vulnerabilities framework

Animoto Account Database Theft (Unattributed, 2018)

2018 · United States · J - Information and communication · Account database theft, data later resold in a dark-web compilation

Analysed 2026-11-07 by Hermes

Open this case in the tool →

What happened

Account database theft, data later resold in a dark-web compilation

Dual vulnerability type

Expertise to Execution

Evidence coverage: 1/6 stages, weight 6/60

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

Why this classification

(1) Strategic expectation: a cloud video-creation service holding personal data for roughly 22 million accounts is expected to be able to tell what left its user database when that database is queried from outside - at minimum to scope an incident accurately for its users and its regulator. (2) In practice: on July 10, 2018 Animoto's own alert caught queries against the user database and they were stopped immediately, but even a forensic investigation with outside experts could not confirm that any data had been removed or whose data was affected, so all users were notified out of an abundance of caution; seven months later, in February 2019, the Register established that a database of 25,402,283 Animoto account records taken in 2018 was being sold on Dream Market - the data had in fact left, and no egress control had detected or interrupted it. The classification rests on the egress stage; the entry vector was never disclosed and is not load-bearing, and detection, containment and disclosure were prompt and self-initiated. (3) That is a knowing-doing gap inside the material column: bulk-export detection (egress/DLP monitoring of the user store) and forensic logging sufficient to reconstruct a query session are known, available technical practice, yet neither delivered here - the remedy existed at expert level and was not realised in operations; no source documents a leadership decision, a deferred programme or an ignored expert warning that would move the break to the decision level, and the company's own statement only concedes it is examining ways to enhance threat detection after the fact. RM33 (Expertise to Execution lens): the specialists could not reconstruct the incident because forensic evidence was inadequate - Animoto stated it could not confirm that data was removed or that any particular user's information was affected, so the scope of the theft remained undetermined until the seller proved it (causal role: Obscuring). RM35 (Expertise to Execution lens): 25.4 million account records left the network undetected and uninterrupted even though the queries themselves were caught the same day - egress visibility over the user database, which a service of this data scale should have had in place, existed neither as tooling nor as tuned monitoring (causal role: Amplifying).

Evidence by stage

Risk markers

Sources

Notes

The row's description (theft of 617 million account details from 16 hacked websites, for sale on the dark web) is the February 2019 Dream Market compilation; The Register confirmed the Animoto lot (25,402,283 accounts, taken 2018) matches Animoto's own July 2018 disclosure, so the row is classified as that 2018 breach with the 2019 resale as egress evidence - description and stated victim are consistent. Entry vector never disclosed (the seller's claim of web-app exploitation is an unverified attacker statement, not coded). Passwords were hashed and salted; whether the salt key was taken stayed unknown, but no source treats the hashing as a failing, so RM26 was not coded. Scale (about 22 million users all notified) per TechCrunch interview with the CEO. No legal or regulatory finding on the victim's controls was located; the case rests on the victim's own notification and the 2019 resale reporting. Analysis under the unattended-run instruction: nothing required consent beyond public source reading.