text
stringlengths 4
429
|
---|
ditional countries considered governed by |
authoritarian of this research deals with attacks carried out by governments |
regimes. |
We believe that activists and journalists in such against opposition and activist groups operating within, as well |
countries may experience harassment or consequences to as outside their borders. One of the most notable cases is |
life or liberty from government surveillance. GhostNet, a large-scale cyber espionage campaign against the |
Finally, we do not explore potential defenses appropriate for Tibetan independence movement [24, 25]. Other work avoids |
protecting the target population in this work. We believe that to drawing conclusions about the attackers [26]. |
2 |
23rd USENIX Security Symposium |
USENIX Association |
Country Date Range Range of Targets Number and Type of Samples Distinct Malware C&C |
Bahrain 4/9/12 |
12 activists, dissidents, trade unionists, 8 FinSpy samples, 7 IP spy links received via private 4 distinct IP addresses |
7/31/13 human rights campaigners, and journalists message, > 200 IP spy links observed publicly |
Syria 2011 to present 10 |
20 individuals with technical back- 40 |
50: predominantly BlackShades, DarkComet, 160 distinct IP addresses |
grounds who receive suspect files from their Xtreme RAT, njRAT, ShadowTech RAT. |
contacts |
UAE 7/23/12 |
7 activists, human rights campaigners, and 31 distinct malware samples spanning 7 types; 5 dis- 12 distinct IP addresses |
7/31/13 journalists tinct exploits |
Table 1: Range of data for the study. |
Country Possible Impacts Probable Impacts |
Bahrain 1. 3 individuals arrested, sen- 1. Activist serving 1 yr in |
tenced to 1 |
12 mo in prison prison |
2. Union leader questioned by 2. Police raid on house |
police; fired |
Syria 1. Sensitive opposition com- 1. Opposition members dis- |
munications exposed to gov- credited by publishing embar- |
ernment rassing materials |
2. Exfiltrated material used to 2. Exfiltrated materials used |
identify and detain activists during interrogation by secu- |
rity services |
UAE Contacts targeted via mal- Password stolen, e-mail |
ware downloaded |
Table 2: Negative outcomes plausibly or quite likely aris- |
ing from attacks analyzed. Figure 1: E-mail containing FinSpy. |
4.1 Bahrain |
3 Data Overview and Implications |
We have analyzed two attack campaigns in the context of |
Our study is based on extensive analysis of malicious files and Bahrain, where the government has been pursuing a crackdown |
suspect communications relevant to the activities of targeted against an Arab-Spring inspired uprising since 2/14/2011. |
groups in Bahrain, Syria, and the UAE, as documented in Ta- The first involved malicious e-mails containing FinSpy, a |
ble 1. A number of the attacks had significant real-world impli- |
lawful intercept |
trojan sold exclusively to governments. The |
cations, per Table 2. In many cases, we keep our descriptions second involved specially crafted IP spy links and e-mails de- |
somewhat imprecise to avoid potential leakage of target identi- signed to reveal the IP addresses of operators of pseudonymous |
ties. accounts. Some individuals who apparently clicked on these |
We began our work when contacted by individuals con- links were later arrested, including Ali (cf. |
1), whose click |
cerned that a government might have targeted them for cyber- appears to have been used against him in court. While both |
attacks. As we became more acquainted with the targeted com- campaigns point back to the government, we have not as yet |
munities, in some cases we contacted targeted groups directly; identified overlap between the campaigns; targets of FinSpy |
in others, we reached out to individuals with connections to tar- appeared to reside mainly outside Bahrain, whereas the IP spy |
geted groups, who allowed us to examine their communications links targeted those mainly inside the country. We examine |
with the groups. For Bahrain and Syria, the work encompassed each campaign in turn. |
10,000s of e-mails and instant messages. For the UAE, the vol- |
ume is several thousand communications. |
FinSpy Campaign. Beginning in April 2012, the authors |
received 5 suspicious e-mails from US and UK-based activists |
4 Case Studies: Three Countries and journalists working on Bahrain. We found that some |
of the attachments contained a PE (.exe) file designed to |
This following sections outline recent targeted hacking cam- appear as an image. Their filenames contained a Uni- |
paigns in Bahrain, Syria and the UAE. These cases have a com- code right-to-left override (RLO) character, causing Windows |
mon theme: attacks against targets |
computers and devices with to render a filename such as gpj.1bajaR.exe instead as |
malicious files and links. In some cases the attackers employed exe.Rajab1.jpg. |
expensive and |
government exclusive |
malware, while in other The other .rar files contained a Word document with an |
cases, attackers used cheap and readily available RATs. Across embedded ASCII-encoded PE file containing a custom macro |
these cases we find that clever social engineering often plays set to automatically run upon document startup. Under default |
a central role, which is strong evidence of a well-informed ad- security settings, Office disables all unsigned macros, so that |
versary. We also, however, frequently find technical and op- a user who opens the document will only see an informational |
erational errors by the attackers that enable us to link attacks message that the macro has been disabled. Thus, this attack was |
to governments. In general, the attacks we find are not well- apparently designed with the belief or hope that targets would |
detected by anti-virus programs. have reduced security settings. |
3 |
USENIX Association |
23rd USENIX Security Symposium |
513 |
Identification as FinSpy: By running the sample using ated with another government [4]. However, a proxy would |
Windows Virtual PC, we found the following string in mem- show gaps in a global IPID as it forwarded traffic; our frequent |
ory: y:\lsvn_branches\finspyv4.01\finspyv2\. observation of strictly consecutive IPIDs thus contradicts this |
This string suggests FinSpy, a product of Gamma Inter- statement. |
national [27]. The executables used virtualized obfusca- Exploitation of captured data: Since we suspected the spy- |
tion [28], which appeared to be custom-designed. We de- ware operator would likely seek to exploit captured credentials, |
vised a fingerprint for the obfuscater and located a structurally particularly those associated with Bahraini activist organiza- |
similar executable by searching a large malware database. tions, we worked with Bahrain Watch, an activist organization |
This executable contained a similar string, except it identi- inside Bahrain. Bahrain Watch established a fake login page |
fied itself as FinSpy v3.00, and attempted to connect to on their website and provided us with a username and pass- |
tiger.gamma-international.de, a domain registered word. From a clean VM, we logged in using these credentials, |
to Gamma International GmbH. saving the password in Mozilla Firefox. We then infected the |
Analysis of capabilities: We found that the spyware has VM with FinSpy and allowed it to connect to the Bahrain C&C |
a modular design, and can download additional modules from server. Bahrain Watch |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.