{ "title": "User-agents: Why and How and How to", "type": "REPORTED", "description": "A TOP SECRET NSA training presentation from 2010 teaching analysts how to use browser user-agent strings — the identifying information your web browser automatically sends to every website you visit — to track surveillance targets across the internet, link their known email activity to their anonymous web browsing, and identify opportunities for hacking their computers based on their operating system and software versions.", "long_description": "This is a TOP SECRET//COMINT//REL TO USA, FVEY NSA training presentation from S2I61 dated July 2010. The presentation teaches analysts how to exploit HTTP user-agent strings for surveillance purposes. User-agent strings are automatically transmitted by web browsers to every website visited, containing the browser type and version, operating system, security settings, installed software frameworks and language preferences. The presentation explains that user-agents can link a target's \"selected\" activity (communications already being collected under tasking) to their \"unselected\" web browsing (untasked internet activity captured in bulk), create tentative links between targets sharing the same user-agent configuration, and identify CNE (hacking) opportunities by revealing a target's operating system version and software configuration. The presentation includes a lighthearted history of browser user-agent spoofing dating back to Internet Explorer pretending to be Mozilla/Netscape to receive frames-enabled web content.\n\nThe practical section demonstrates querying user-agents in XKeyscore and PINWALE. The key analytical judgment taught is assessing user-agent reliability — when an IP address shows only two user-agent strings with low traffic volume, the correlation between user-agent and target is \"probably reliable.\" When the same IP shows many different user-agent variations (the example shows 17 distinct strings including variations in .NET CLR versions, Google Toolbar versions and InfoPath installations), reliability \"plummets\" and cookie/TDI correlation becomes necessary. The presentation warns against assuming complex user-agents are unique — a standard Chrome 6 on Windows 7 string applies to every user with that configuration. Mobile user-agents are highlighted as particularly valuable because they typically include the phone model, creating \"IMEI correlation opportunities\" for linking web activity to specific physical handsets.\n\nThe presentation concludes with the operational caveat that user-agents can be manipulated or removed entirely — via Firefox plugins, Opera browser settings or the TOR Button — allowing targets to present a different user-agent for each browsing session. The final slides advertise S2I61's broader services including help with XKeyscore fingerprint development, target technology analysis and collateral description development, positioning the unit as an internal SIGDEV technical support team.", "tags": [ "User-Agent", "HTTP", "browser fingerprinting", "TDI", "cookies", "Computer Network Exploitation", "operating system", "Windows", "Firefox", "Chrome", "Safari", "Opera", "Internet Explorer", "mobile user-agent", "iPhone", "Android", "BlackBerry", "IMEI", "Gecko", "AppleWebKit", "Presto", "encryption", "USSID-18", "compliance", "web browsing", "SIGDEV", "target tracking" ], "document": { "creator": "NSA", "date_created": "2010-07-01", "classification": [ "TOP SECRET", "COMINT" ], "dissemination": [ "REL TO USA, FVEY" ], "banner": "TOP SECRET//COMINT//REL TO USA, FVEY", "genre": "slides", "programs": [ "XKEYSCORE", "PINWALE", "AGILITY" ], "countries": [], "people": [] }, "publication": [ { "publisher": "The Intercept", "date_published": "2015-07-01", "title": "XKEYSCORE: NSA’s Google for the World’s Private Communications", "link": "https://theintercept.com/2015/07/01/nsas-google-worlds-private-communications/", "reporters": [ "Marquis-Boire, Morgan", "Greenwald, Glenn", "Lee, Micah" ], "contributors": [], "doc_source": "https://theintercept.com/document/2015/07/01/user-agents/", "doc_url": "https://s3.documentcloud.org/documents/2116232/user-agents.pdf" } ], "pdf_sha256": "0177d72ec9aa3f502ed8c042c1c778fd558894c8757f18610bf84ca07a63790c", "pages": 42 }