← Breach analysis by the Dual Vulnerabilities framework

Booz Allen Hamilton Military Data Leak (Anonymous, 2011)

2011 · United States · M - Professional, scientific and technical activities · Data theft via insecurely protected server; public data dump

Analysed 2026-11-07 by Hermes

Open this case in the tool →

What happened

Data theft via insecurely protected server; public data dump

Dual vulnerability type

Expertise to Execution

Evidence coverage: 3/6 stages, weight 17/60

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

Why this classification

(1) Booz Allen Hamilton ran systems for US government clients - its own press release confirms it was running a learning management system for a government agency - and as a leading defense and intelligence contractor the expectation was that military account data held on systems it operated was protected differently from commodity data (strong, salted credential storage, segregated sensitive datasets) and that breaches would be disclosed to clients and affected people. (2) On 11 July 2011 Anonymous/AntiSec published a database of roughly 90,000 military e-mail addresses with password hashes plus data from the Defense Acquisition University training system; independent inspection of the dump (Ars Technica: SHA-1 hashes; SecurityWeek: Base64 hashes) showed credentials stored in weakly protected form in a single, bulk-dumpable database, and the full SQL dump left the network undisturbed. The company declined comment on day one citing its policy, confirmed on day two with a scope-limiting preliminary statement, and never disclosed the entry vector - the only account of it (a server that 'basically had no security measures in place') is the attackers' unverified claim. (3) The classification rests on the egress stage, the only load-bearing gap documented by sources other than the attacker: the knowledge that stored credentials must be strongly hashed and sensitive datasets segregated sat squarely with the contractor's own specialists (material origin), yet the data lay in bulk-dumpable, weakly protected form in an architecture the company ran for clients (execution break) - no source documents a leadership decision that caused this, so it is not a leadership break. The impact, mass publication of military account data, was directly produced by known remedies not executed - a knowing-doing gap, not merely weak security. Detection and recovery show no documented gap; the entry vector is not load-bearing for the type because it rests on speculation alone; the response gap (doctrinal silence policy, disclosure forced by the leak, doctrinal + leadership = Strategy to Leadership) is contributing and is the runner-up. RM26 (Expertise to Execution lens): the credential data at rest was protected only by weakly hashed storage - SHA-1 verified in the published dump, non-salted MD5 per the attackers - so the specialist remedy of properly protected password storage was never realised on the server, amplifying the exposure of 90,000 accounts (causal role: Amplifying). RM16 (Expertise to Execution lens): one contractor database co-mingled ~90,000 military account records with personal attributes and data of another agency's training system, so data ownership and lineage were not operationalised and more sensitive data than the training function required was exposed in one dump (causal role: Amplifying). RM39 (Expertise to Execution lens): scope could not be established quickly - the company's only statement was 'preliminary' ('At this time, we do not believe that the attack extended beyond data pertaining to a learning management system for a government agency', a full review pending) - so disclosure was forced and framed by the attacker's leak while individuals learned of the exposure from the dump (causal role: Constraining).

Evidence by stage

Risk markers

Sources

Notes

Link check: investors.boozallen.com blocks automated access (HTTP 403); the page was opened and read during research. EuRepoC row; the case file's source_url (Forbes, Andy Greenberg) returned HTTP 403 and has no web.archive.org snapshot, so it is not cited and the analysis rests on the sources listed above, all opened in this run. Entry stage rests solely on Anonymous's unverified claim (reported by SecurityWeek/BBC); Booz Allen never disclosed the vector - entry is therefore speculation-sourced and not load-bearing for the DV-Type. Hash reports conflict: Anonymous claimed non-salted MD5, Ars Technica verified SHA-1, SecurityWeek reported Base64 hashes and corrected the MD5 claim; RM26 is coded on the independent dump verification. Disclosure stage per eurepoc_extra: incident disclosed by attacker; company statement followed on 12-07-2011. 2011 reporting leaves detection timing, dwell time and remediation steps undocumented; no regulator or congressional record specific to this incident was found. No related_rows in the case file and no existing result for this victim in the output directory.