first commit

This commit is contained in:
libroot 2026-11-19 04:39:11 +02:00
commit 0d90363116
Signed by: libroot
GPG key ID: 2AC68C029EBE9ACE
6510 changed files with 679146 additions and 0 deletions

View file

@ -0,0 +1,73 @@
{
"title": "CNE Presence in CT10 Status Report",
"type": "REPORTED",
"description": "A TOP SECRET NSA status report from April 2008 describing how a counterterrorism analysis team began importing data from the agency's hacking operations into their tracking system — specifically logs from FOXACID, an NSA tool that infects targets' computers mainly through their web browsers — enabling real-time alerts when new targets were successfully hacked, with plans to expand into tracking four named surveillance implants (VALIDATOR, UNITEDRAKE, OLYMPUS and STRAITBIZARRE) as they phone home to NSA servers for instructions and stolen data.",
"long_description": "This is a TOP SECRET//COMINT//NOFORN status report dated 1 April 2008, authored by a SNIP intern and a first lieutenant, describing work completed on integrating NSA Tailored Access Operations CNE data into the CT10 pilot project. CT10 appears to be a counterterrorism analytical platform designed to fuse multiple data types — DNR (telephony), DNI (internet) and CNE (hacking) — into a unified analytical environment.\n\nThe report describes two parallel efforts. The first was acquiring and processing active SIGINT from TAO's hacking operations into CT10 for real-time alerting. The second was designing an extension to CT10's data model to support deeper analytical exploitation of CNE data. The primary customer driving requirements was GHOSTWolf — an internal NSA team or mission whose specific focus is not described but whose requirements shaped the prioritisation of data flows.\n\nThe immediate achievement reported is the ingestion of FOXACID log files into CT10, established in mid-February 2008. FOXACID is NSA's content-based exploit delivery system — a server infrastructure that delivers tailored exploits to targets' web browsers when they visit specific URLs. The log files record each exploitation attempt and its outcome. CT10 mapped the flow of FOXACID data through NSA and established a feed from Protocol Exploitation. The existing CT10 data model proved flexible enough to accommodate FOXACID events by defining a new event type and populating events with selectors extracted from each FOXACID log entry. A new selector type was created for implant IDs, allowing them to be added to watch lists and target selectors so analysts could track all events associated with specific implants. The system was alerting on FOXACID exploitation activity through geolocation-based tipping in GeoT, IRC and email alerting through Agent Logic, and the iSpace dashboard.\n\nThe near-term roadmap describes the next planned data flows in priority order. First was implant callback data — allowing analysts to track when deployed implants communicate back to listening posts for tasking and exfiltration. Four specific implant families are named: VALIDATOR, OLYMPUS (spelled OLYUMPUS in the document), UNITEDRAKE and STRAITBIZARRE (likely STRAITBIZZARE/STRAITBIZARRE). These represent different tiers of TAO's implant toolkit — VALIDATOR is a lightweight initial-access implant, UNITEDRAKE is a more capable Windows implant with modular plug-in architecture, and STRAITBIZARRE is an implant deployed on compromised internet hosts as part of the QUANTUM shooter infrastructure. After callback data, the plan was to ingest collected data from the implants themselves — exfiltrated files and plug-in results — giving CT10 analysts a complete picture from initial exploitation through data theft.\n\nAdditional CNE datasets outside TAO's direct production were also identified as valuable. GOLLUM is described as a \"partner implant\" — meaning it was developed or operated by a Five Eyes partner agency — which GHOSTWolf wanted ingested into CT10. RADIUS logs — ISP dial-up customer authentication records — were identified as an excellent bridging dataset creating natural links between DNR and DNI data by connecting phone numbers used for dial-up access to internet activity.\n\nThe challenges section reveals significant organisational and classification issues within TAO's data infrastructure. The complexity of TAO's internal data flows made locating the ultimate data owner for each dataset difficult, and CT10 was not embedded in TAO's requirements generation process, limiting its ability to establish new data feeds. Most significantly, FOXACID log files carried no native classification markings, and compartmented target data and FISA-authorised collection had to be manually stripped by TAO before data could be shared with external consumers like CT10. The report explicitly identifies the lack of universal marking and classification of CNE data as a systemic problem affecting all consumers of TAO data, not just CT10. TAO engineers were developing an RDF (Resource Description Framework) database system intended to centralise their data and make it available to external customers, which would address many of these issues.",
"tags": [
"Tailored Access Operations",
"CT10",
"Computer Network Exploitation",
"Agent Logic",
"iSpace",
"protocol exploitation",
"implants",
"callback",
"listening post",
"exfiltration",
"RADIUS",
"ISP logs",
"FISA",
"classification markings",
"DNR",
"DNI",
"SNIP intern"
],
"document": {
"creator": "NSA",
"date_created": "",
"classification": [
"TOP SECRET",
"COMINT"
],
"dissemination": [
"NOFORN"
],
"banner": "TOP SECRET//COMINT//NOFORN//20291123",
"genre": "report",
"programs": [
"FOXACID",
"GHOSTWOLF",
"VALIDATOR",
"OLYUMPUS",
"UNITEDRAKE",
"STRAITBIZZARE",
"GEOT",
"RADIUS"
],
"countries": [],
"people": []
},
"publication": [
{
"publisher": "Der Spiegel",
"date_published": "2015-01-17",
"title": "The Digital Arms Race: NSA Preps America for Future Battle",
"link": "https://www.spiegel.de/international/world/new-snowden-docs-indicate-scope-of-nsa-preparations-for-cyber-battle-a-1013409.html",
"reporters": [
"Appelbaum, Jacob",
"Gibson, Aaron",
"Guarnieri, Claudio",
"Müller-Maguhn, Andy",
"Poitras, Laura",
"Rosenbach, Marcel",
"Ryge, Leif",
"Schmundt, Hilmar",
"Sontheimer, Michael"
],
"contributors": [],
"doc_url": "https://cdn.prod.www.spiegel.de/media/3a84ee44-0001-0014-0000-000000035659/media-35659.pdf"
}
],
"pdf_sha256": "f4f9171fb6c6273499bc75406c3fe348efac0a45ec3e676406454233dbb5568e",
"pages": 4
}

View file

@ -0,0 +1,68 @@
TOP SECRET//COMINT//NOFORN//20291123
CNE Presence in CT10 Status Report
Authors:
[REDACTED], SNIP Intern
[REDACTED], 1LT
(U//FOUO) Executive Summary
(S//SI//REL USA, FVEY) This paper discusses the work completed as of 1 April 2008 on the inclusion of CNE data into the CT10 Pilot Project. CT10 engaged two fronts in this effort. First, was the acquisition and alerting of active SIGINT into CT10. Second, was the design of an extension to the current CT10 data model to support further analytic development of CNE data. As a result of this work, CT10 now supports alerting of the establishment of new TAO points of presence on the Internet with respect to certain exploitation methodologies. Additionally, this paved the road with internal and external partners for the continued development of the CT10 CNE data model extension.
(S//SI//REL USA, FVEY) Specific conclusions and recommendations are contained at the close of this paper. However, three key points underpin a majority of the information in this paper:
* (U//FOUO) Priority support to CT10 data acquisition efforts.
* (U//FOUO) Further building a DNI/CNE analytic and development team.
* (S//SI//REL USA, FVEY) Development of a strategic vision for use of CNE data in CT10.
(U//FOUO) TAO Data Flow Acquisition
(S//SI//REL USA, FVEY) The first step in processing CNE data is to identify and acquire CNE data flows in order to route this information into CT10. This work presents its own unique challenges,, as described in the next section.
(U) Current State
Derived From: NSA/CSSM 1-52
Dated: 20070108
Declassify On: 20291123
TOP SECRET//COMINT//NOFORN//20291123SECRET//COMINT//REL TO USA,
FVEY//20291123
TOP SECRET//COMINT//NOFORN//20291123
(S//SI//REL USA, FVEY) CT10 is ingesting a single flow of CNE data, specifically FOXACID log files. This was identified as a top priority by CT10 customers GHOSTWolf, therefore it was chosen to be the first CNE data flow to pursue. After mapping the flow of FOXACID data throughout NSA, CT10 established a flow of this data from Protocol Exploitation starting in the middle of February 2008.
(S//SI//REL USA, FVEY) Because the CNE data model extension had not been implemented at this time, the only way to process FOXACID logs was to fit the information into the current CT10 data model. This was possible primarily because of the natural flexibility in the current model. It was easy to define a new "event type" within the CT10 ingestion system, and populate events with all selectors pulled from the each FOXACID event. Another benefit of the current model design was that it became easy to define a new selector type for an implant ID and easily integrate it into the system. This allows implant IDs to be added to watch lists and as target selectors, so that CT10 users can track any events that are associated with a defined set of implants. Therefore, if CT10 acquired a data flow containing implant callback information, it would be easy to add this into the current data model, giving analysts additional capability to follow implants deployed through FOXACID..
(S//SI//REL USA, FVEY) As a result of this work, the CT10 system is now alerting on FOXACID exploitation activity through a suite of alerting technologies. This includes geo-locational tipping in GeoT, and alerting through Agent Logic (IRC and e-mails) and the iSpace dashboard.
(S//SI//REL USA, FVEY) One conclusion reached from developing the FOXACID ingestion code is that the current CT10 data model can support more CNE data than previously believed. As new CNE flows are introduced into the system, it may be possible to easily add them to the current data model if the extension is still not implemented.
(U//FOUO) Near and Long Term Goals for TAO Data Flow Acquisition
(S//SI//REL USA, FVEY) The near term goal regarding CNE data flow into CT10 is to identify additional CNE data flows that build off of the FOXACID data. Based on the requirements of GHOSTWolf, the immediate next step would be to acquire implant callback data flows. This would allow analysts to track when their implants are calling back to listening posts for tasking and exfiltration. The implants in question are VALIDATOR, OLYUMPUS, UNITEDRAKE, and STRAITBIZZARE.
TOP SECRET//COMINT//NOFORN//20291123
TOP SECRET//COMINT//NOFORN//20291123
(S//SI//REL USA, FVEY) After establishing this data flow, the next step would be to ingest collected data from the implants. This includes any exfiltrated files, and implant plug-in results. With each new data flow, CT10 analysts will have a more complete picture of CNE activity related to their targets. However, these types of data do not naturally fit into the current data model, which is why the proposed extension is so important.
(S//SI//REL USA, FVEY) Additionally, there are other CNE datasets not created by TAO that would be useful to CT analysts. GOLLUM, a partner implant, is one such data set that GHOSTWolf has indicated that would be useful to ingest into CT10. RADIUS logs (ISP dial up customer records) are also an excellent source of information, and create a natural link between DNR and DNI datasets.
(U) Challenges / Way Ahead
(S//SI//REL USA, FVEY) There are a number of challenges for securing further CNE data from TAO. First, given the complexity of TAO data flow, locating the ultimate data owner is challenging. Though everyone in TAO CT10 has interacted with on data flow issues has been extremely helpful, no new CNE flows have been established. This is largely due to the fact that CT10 is not imbedded within the requirements generation process in TAO. Additional transparency on TAO data flow requirements will greatly ease this issue.
(S//SI//REL USA, FVEY) One other issue that would greatly help other systems in using TAO data, is a universal marking and classification of CNE data. In the case of FOXACID, there are no native classification markings on the files, and compartmented targets as well as FISA data have to be stripped out by TAO before data is sent on to external customers such as CT10. As CT10 becomes PL3 accredited, it becomes critical that CT10 ingests full FOXACID records. Unfortunately, the lack of a unique marking and classification for each file creates additional challenges that complicate the oversight, classification and policy surrounding CNE data. Furthermore, additional discussions with TAO have indicated that this may be an issue for other data flows as well. The addition of classification markings to TAO data feeds would greatly assist all consumers of TAO data with maintaining the proper security procedures.
(U//FOUO) TAO engineers have worked towards creating an RDF database system that would house all of their information, and make it available to external customers. As this capability is developed, CT10 could benefit greatly from becoming a consumer of this data. This system would address a number of issues, including the marking and classification challenge addressed in the previous paragraph.
TOP SECRET//COMINT//NOFORN//20291123
TOP SECRET//COMINT//NOFORN//20291123
(S//SI//REL USA, FVEY) In summary, the way ahead is to address these existing challenges in order to effectively push more TAO data flows out to analysts in near real time. CT10 should also remained synched with TAO's effort to develop and deploy the RDF database. Finally, CT10 needs to implement the data model extension and move any currently ingested TAO data into the new model.
TOP SECRET//COMINT//NOFORN//20291123