94 lines
5 KiB
JSON
94 lines
5 KiB
JSON
{
|
|
"title": "SKYNET: Applying Advanced Cloud-based Behavior Analytics",
|
|
"type": "REPORTED",
|
|
"description": "A classified NSA SIGINT Development presentation detailing SKYNET, a cloud-based machine learning system that analyzes bulk Pakistani call data records to identify suspected couriers and terrorists through automated travel pattern analysis, behavioral profiling, and social network mapping — with the goal of supporting targeting operations against groups including the Haqqani Network.\n",
|
|
"long_description": "This top secret NSA SIGINT Development presentation describes SKYNET, a collaborative cloud analytics project spanning five NSA organizations across three directorates — Signals Intelligence, Research, and Technology — with external partnerships including MIT Lincoln Labs and Harvard University. The system's core function is the application of geospatial, geotemporal, pattern-of-life, and travel analytics to bulk call data records (CDRs) acquired from major Pakistani telecommunications providers through a dataflow called DEMONSPIT. Raw CDRs are normalized through TUSKATTIRE and distributed across multiple cloud systems including GMHalo, GMPlace, Cloud 14, and Bulldozer/MDR2, with processed data flowing into analyst-facing tools including MAINWAY, FASCIA, JEMA, CINEPLEX, SMARTTRACKER, and SORTINGLEAD.\n\nThe analytical framework SKYNET applies is built around a set of behavioral indicators associated with courier and terrorist activity: regular travel between specific cities on fixed schedules, frequent SIM or handset swapping, excessive phone power-downs, low usage with incoming calls only, co-traveler identification, and proximity to known tasked selectors. The document presents specific operational use cases focused on the Peshawar-Lahore-Faisalabad corridor and courier routes between Afghanistan and Pakistan, including a map of a described courier path spanning provinces from Baghlan and Konarha through Kabul, Jalalabad, Peshawar, Lahore, and into Waziristan. Sample travel reports show SKYNET applied against Haqqani Network-associated selectors, with results including IMSI numbers, contact counts, selector swapping behavior, and travel phrases across Afghanistan, Pakistan, Iran, Bahrain, and the UAE.\n\nThe triage workflow integrates three tools sequentially: SMARTTRACKER identifies co-location coincidences between suspects across time and cell tower locations; RT-RG analytics identify meeting partners and traveling pairs (\"sidekicks\") at destination cities; and JEMA serves as the integrating human-in-the-loop environment where analysts evaluate travel reports, prioritize selectors, and generate targeting leads. The document includes a coincidence report showing one redacted target co-located with others at up to 24 distinct locations 101 times — illustrating the granularity of behavioral mapping the system enables against individuals who have not yet been formally tasked as targets.",
|
|
"tags": [
|
|
"pattern-of-life",
|
|
"travel analytics",
|
|
"call detail records",
|
|
"bulk collection",
|
|
"Haqqani Network",
|
|
"courier tracking",
|
|
"machine learning",
|
|
"cloud analytics",
|
|
"behavioral profiling",
|
|
"SIM swapping",
|
|
"GSM",
|
|
"handset swapping",
|
|
"geolocation",
|
|
"cell towers",
|
|
"IMSI",
|
|
"MSISDN",
|
|
"counterterrorism",
|
|
"targeting",
|
|
"social network analysis",
|
|
"co-travelers",
|
|
"metadata analysis"
|
|
],
|
|
"document": {
|
|
"creator": "NSA",
|
|
"date_created": "",
|
|
"classification": [
|
|
"TOP SECRET",
|
|
"COMINT"
|
|
],
|
|
"dissemination": [
|
|
"REL TO USA, AUS, CAN, GBR, NZL"
|
|
],
|
|
"banner": "TOP SECRET//COMINT//REL TO USA, AUS, CAN, GBR, NZL",
|
|
"genre": "slides",
|
|
"programs": [
|
|
"SKYNET",
|
|
"DEMONSPIT",
|
|
"TMAC",
|
|
"FASTSCOPE",
|
|
"TUSKATTIRE",
|
|
"GMHALO",
|
|
"GMPLACE",
|
|
"BULLDOZER",
|
|
"MDR2",
|
|
"CINEPLEX",
|
|
"JEMA",
|
|
"FASCIA",
|
|
"CLOUD",
|
|
"DPS",
|
|
"SEDB",
|
|
"SMARTTRACKER",
|
|
"SORTINGLEAD",
|
|
"ASSOCIATION",
|
|
"BANYAN",
|
|
"SMARTTRACKER",
|
|
"RT-RG",
|
|
"SNOWHAZE",
|
|
"MASTERSHAKE",
|
|
"MAINWAY"
|
|
],
|
|
"countries": [
|
|
"Afghanistan",
|
|
"Pakistan",
|
|
"Iran",
|
|
"Bahrain",
|
|
"United Arab Emirates"
|
|
],
|
|
"people": []
|
|
},
|
|
"publication": [
|
|
{
|
|
"publisher": "The Intercept",
|
|
"date_published": "2015-05-08",
|
|
"title": "U.S. Government Designated Prominent Al Jazeera Journalist as \"Member of Al Qaeda\"",
|
|
"link": "https://theintercept.com/2015/05/08/u-s-government-designated-prominent-al-jazeera-journalist-al-qaeda-member-put-watch-list/",
|
|
"reporters": [
|
|
"Currier, Cora",
|
|
"Greenwald, Glenn",
|
|
"Fishman, Andrew"
|
|
],
|
|
"contributors": [],
|
|
"doc_source": "https://theintercept.com/document/2015/05/08/skynet-applying-advanced-cloud-based-behavior-analytics/",
|
|
"doc_url": "https://s3.documentcloud.org/documents/2074625/skynet-applying-advanced-cloud-based-behavior.pdf"
|
|
}
|
|
],
|
|
"pdf_sha256": "d3699e74af2ba8c621452d3dcc07bacdf8b0846060fc020793b8b364dd330b59",
|
|
"pages": 20
|
|
}
|