Dual-use Research in Ransomware Attacks:
A Discussion on Ransomware Defence Intelligence
Ziya Alper Genc¸
a
and Gabriele Lenzini
b
Interdisciplinary Centre for Security, Reliability and Trust (SnT), University of Luxembourg, Luxembourg
Keywords:
Ransomware, Double Use Research in Cryptography, Threat Intelligence and Counter-intelligence.
Abstract:
Previous research has shown that developers rely on public platforms and repositories to produce functional but
insecure code. We looked into the matter for ransomware, enquiring whether also ransomware engineers re-use
the work of others and produce insecure code. By methodically reverse-engineering 128 malware executables,
we have found that, out of 21 ransomware samples, 9 contain copy-paste code from public resources. Thanks
to this finding, we managed to retrieve the decryption keys with which to nullify the ransomware attacks.
From this fact, we recall critical cases of code disclosure in the recent history of ransomware and, arguing
that ransomware are components in cyber-weapons, reflect on the dual-use nature of this research. We further
discuss benefits and limits of using cyber-intelligence and counter-intelligence strategies that could be used
against this threat.
1 INTRODUCTION
In anti-ransomware research, ransomware samples
are routinely analyzed. The goal is to understand how
they generate or retrieve the encryption keys; how
they search, sort and prioritize which files to target
first; and which files they encrypt first and by us-
ing which encryption algorithm. In this quite me-
thodical work, it is routine to reverse engineer ran-
somware samples and analyze their source codes.
While performing this task, we found that some piece
of code was not original but copy-and-pasted from
well-known public repositories or developpers com-
munities. From this discovery, with some additional
work, we managed to build a decryptor for those ran-
somware samples.
Although our discovery is not surprising—
researchers have already commented on how codes
from public repositories is re-used and how this im-
pacts security (e.g., see (Fischer et al., 2017))—
realizing that also ransomware’s security depends on
public code has captured our attention. We started
wondering whether there were other cases of copy-
and-pasted code in ransomware. And we started re-
flecting on which consequences such re-use of code
may bring into the fight against ransomware attacks.
a
https://orcid.org/0000-0001-7198-7437
b
https://orcid.org/0000-0001-8229-3270
This articles report on our insights on the subject.
Although motivated by some experimental find-
ings, our contribution is purely argumentative. But,
by developing our argument rigorously, we hope to
contribute to a scientific discussions on “the matter”.
And being “the matter” related to dual-use of con-
cern in ransomware research, we intend to embark on
other questions as well: What famous precedents ex-
ist in the recent history of ransomware that could en-
lighten us on the pros and cons of dual-use research?
Should ransomware be considered components of a
cyber-weapon? And, as such, are there reasons to
classify ransomware as having military use? Thus,
would it be reasonable to resort to intelligence and
counter-intelligence strategies, such as those suggest-
ing to contain information spreading in case of an at-
tack or to control public information, to mitigate the
threat? We restrict our argument to cryptographic ran-
somware, those which rely on cryptography. Other
kind of ransomware, e.g., those which aim to distress
victims to pay up but, like the scareware, only pre-
tend to use encryption but do not, are excluded from
the discussion.
2 PRELIMINARIES
Is copy-and-paste from public repositories a practice
in ransomware engineering? To investigate the ques-
Genç, Z. and Lenzini, G.
Dual-use Research in Ransomware Attacks: A Discussion on Ransomware Defence Intelligence.
DOI: 10.5220/0009000505850592
In Proceedings of the 6th International Conference on Information Systems Security and Privacy (ICISSP 2020), pages 585-592
ISBN: 978-989-758-399-5; ISSN: 2184-4356
Copyright
c
2022 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
585
tion we have first to collect and obtain the code of
real-world ransomware samples and reverse engineer
it.
The most accurate way to accomplish this latter
is to decompile the malicious binaries. The task be-
comes quite practical if the malware is implemented
using the .NET framework. Looking into malicious
.NET assemblies downloaded from “Hybrid Analy-
sis”, an automated malware analysis platform (Hybrid
Analysis, 2019). Hybrid Analysis utilize sandboxing
technique to determine if an executable exhibits ma-
licious behaviour or poses no specific threat. From
it, we collected ransomware samples by searching on
report database with the following settings: (i) Exact
Filetype Description as Mono/.Net assembly, for
MS Windows; (ii) Verdict field as Malicious; and
(iii) Hashtag field as #ransomware.
On a initial set of 128 executable, we applied
dnSpy (dnSpy, 2019), a tool to obtain source codes.
39 samples, obfuscated, precluded any analysis. Of
the remaining 89, we manually perused the source
code, searching for key generation and encryp-
tion routines. 68 samples turned out to be non-
cryptographic ransomware, with no such routines in
their program body. The remaining 21 cryptographic
ransomware samples were our final data set.
Using the found crypto-related code lines (e.g.,
key derivation, encryption) as keywords, we searched
for those lines in public developer platforms. When
analyzing the hits, we compared the semantics of code
snippets, naming of constants and variables, function
signatures, strings, and error messages. From this
searching and matching we discovered that some code
was a verbatim copy-paste. Other code resulted, at
least apparently, a plagiarism of some public available
code.
Were we witnessing code-reuse (i.e., dual-use) in
ransomware? Before claiming code-reuse, we had
to verify whether the code had been published be-
fore the first appearance of it in the malware. There
should also be a reasonable time frame between the
two events. The date of the first appearance of a
ransomware, checked by using VirusTotal (VirusTotal
Threat Intelligence, 2019), has been compared with
the date on which the knowledge was first shared
on online. A double-checked on the integrity of the
pieces of information available on the executable was
also performed. According to our findings, at least
9 out of 21 ransomware samples resulted to contain
snippets bearing a marked resemblance to codes at
online resources, this leading us to conclude that they
are in fact a copy-paste.
In the following, we can comment on an ex-
cerpt from the ransomware samples (see Table1) that
we have found being a copy-paste from (i) a public
repository of fully functional ransomware prototypes;
(ii) tutorials and posts at developer communities. We
also elaborate, where possible, about where the orig-
inal code comes from, and about its cryptographic
qualities.
Ransomware from Repositories of Fully Func-
tional Prototypes. Tiggre, see Table 1, is a sample
of cryptographic ransomware that uses a key gener-
ation function that is copy of a piece of public code
known as HiddenTear (S¸en, 2015b) (see Fig. 2 and
Fig. 1). From it, Tiggre inherits a weakness: the
password is generated using the outputs of a cryp-
tographically weak algorithm. In fact, the same au-
thor of HiddenTear had developed a decryptor by
using this weakness (S¸en, 2015a). We tell the full
story later, but what counts for now is that the open-
source ransomware HiddenTear is a very famous ran-
somware code, which was posted publicly in 2015 al-
legedly for educational purpose. Since then, cyber-
criminals have been using it as a source of inspiration
for their ransomware variants (van der Wiel, 2016).
This was also the case for Tiggre.
The original HiddenTear works as follows: it
generates a password by calling CreatePassword
which is shown in Fig. 1. The password, from which
the encryption keys are derived, is sent to Command
and Control (C&C) server. Next, before notifying the
user, the ransomware attempts to encrypt all the files
in test folder under the user’s Desktop directory.
Ransomware authors that copy from HiddenTear
had to implement their own back-ends before having a
working ransomware, but HiddenTear remains their
point of reference. We have found that basic func-
tionalities such as password generation and encryp-
tion blocks have been replicated from HiddenTear:
for each file, the encryption key is derived from the
same master secret, the password; this latter is gen-
erated using System.Random, a class that provides
(cryptographically weak) pseudo random numbers.
From a cryptographic point of view, the outputs
of System.Random is reproducible when using the
same seed and its secrets are vulnerable to a forensics
analysis. But other variants of HiddenTear eliminate
this weakness: the weak key generation method is not
seen in those samples.
Ransomware from Community Platform. Confi-
dentiality of data is a highly demanded and legitimate
need in the digital world. While cryptographic tech-
niques can be used to protect the secrecy of data, de-
veloping a security application is an error-prone pro-
cess. Therefore, developers who recently entered in
ICISSP 2020 - 6th International Conference on Information Systems Security and Privacy
586
Table 1: SHA256 digests and family names of the samples. To determine the family name, we applied AVCLASS tool on
the labels provided by AV vendors which we obtained from VirusTotal. SINGLETON denotes that the tool could not find any
plurality among the labels for that sample, i.e., no vendor agreed on the family name.
# SHA256 Digest Family Reuse From
1 0e5a696773b0c9ac48310f2cda53b1742a121948df5bcb822f841d387f0f5f68 Jigsaw
2 1d57564398057df99d73cca27015af24142c25828287837c73d2daf0b3c3af5b Mimikatz
3 1ebdbfea6ab13f258a7d00dea47de48261cfb84d52ebbb6f282498c3ab1b1b39 Occamy
4 3b4aaf37510c0f255e238c81b7e1a446bfa925bd54f93969c3155d988fbb6501 HiddenTear (S¸en, 2015b)
5 41ee4623d60544dd0ca16f6177565d99825afb38b932ccecc305ef2fc20e03f4 HiddenTear (S¸ en, 2015b)
6 58d11ef74b062e9996e75d238501a3f4d23691b101997d898d478696795ae3ff CloudSword (den Bosch, 2005)
7 662d0f034f2852e4e43d22a3625c1c8600c3d36660b596db1d6bad5c4980d9df Ryzerlo (S¸en, 2015b)
8 66a3172e0f46d4139cc554c5e2a3a5b6e2179c4a14aff7e788bb9cc98a2219d5 Tiggre
9 7cdd7e30c7091fd2fa3e879dd70087517412a165bf14c4ea4fd354337f22c415 HiddenTear (S¸ en, 2015b)
10 87ce0b2e22b02572146676277cd6e9d89225e75361d1b696555cfe695c2e1f45 SINGLETON
11 894aa842c129b39c0b9a7d575133d68b25de2ecd4e777f29e58481d30dfb6f4e Omegax
12 950be5b5501ee84b1641c3a9a780242a57cdd412892c781eac8781498bf11f3e Bobik
13 951d78dd92eba7daa3ef009ce08bba91a308e13bdeb8325af35bc8202bd76e9b Tiggre
14 b5f3a090556ea30210a23fc90b69c85c68e8e08c89fbe58eb6a829e356dcc42e Occamy
15 ce53233a435923a68a9ca6987f0d6333bb97d5a435b942d20944356ac29df598 Crypren (Lydford, 2008) & (Johnson, 2008)
16 d36e6282363c0f9c05b7b04412d10249323d8b0000f2c25f96c6f9de207eedf8 HiddenTear (S¸ en, 2015b)
17 def09368d22c7b3f6a046ef206a57987095b2f4ddae1d26c6ef2594d6be09bfc Diztakun
18 e2ac9692c0816ccd59d1844048c6238dc5d105b0477620eeb1cdb0909804a787 WhiteRose (S¸en, 2015b)
19 f37080ee4cc445919cae0b1eb40eff46571f7ce0d85b189321d80a41c8752212 SINGLETON
20 f535879cf05a099bf0f6d2a7fa182d399ec9568f131abb23d9fb98418f45789d Perseus (den Bosch, 2005)
21 fd99bfeac78c087a9dc9d4c0c1d26a7ea9780a330f88ba0d803f3464221b4723 SINGLETON
the field of cybersecurity might need to use the help of
online tutorials. For example, Fig. 3 shows a post on
CodeProject website (Lydford, 2008) which explains
a simple way to encrypt a file using a key derived
from a password in C# language (see Fig. 3). The
function, EncryptFile, is poorly written from a cryp-
tographic point of view. There are weaknesses, such
as (i) presence of a hard-coded secret in the code; and
(ii) improper key derivation, to name a few. That said,
we found a Crypren ransomware variant (see Table 1)
which copies the file encryption and decryption func-
tions from (Lydford, 2008). Fig. 4 shows the function
modified by the ransomware author, who disdained
to write a password generation method and even used
almost the same hard-coded secret.
Furthermore, the same Crypren sample contains
the exact code snippet shared at another developer
community (Johnson, 2008). That piece of code, is
meant to impersonate another user, i.e., to launch a
process under that user’s account. However, the said
code portion is not used/referenced by the program.
In another case, we observed that an online tu-
torial published in 2005 inspired two ransomware
samples: Perseus and CloudSword (see Table 1).
The post, available at (den Bosch, 2005), explains
how to encrypt files with a user-supplied password
in VB.NET programming language. Many portions
of the code is reused by the ransomware samples, bar
the part which takes input (i.e., the password) from the
user. Alternatively, the Perseus variant uses an em-
bedded password to derive key, while the CloudSword
variant uses the System.Random to generate a pass-
word from which the key is derived. The CloudSword
sample even contains the exact error messages as in
the full project at (den Bosch, 2005); the Perseus
sample uses the same code portions as in the tutorial,
that without error messages.
Discussion. From our findings, we can conclude
that certain ransomware engineers do copy-and-paste
code from public sites. Surely, this conclusion cannot
be representative of how all ransomware variants are
coded. We do not even know whether who took ad-
vantage the public resources are professionals or am-
ateurs, and it may be inherently hard to investigate for
an answer on this matter due to the difficulty to reach
out ransomware developers. However, we speculate,
ransomware engineers are likely not in a different po-
sition than security developers. In (Acar et al., 2017),
it is reported that in a population of three hundreds de-
velopers among which also professionals, only a quar-
ter relied on the official documentation, while the rest
consulted “the Internet”, inevitably relaying in their
code errors na
¨
ıvities “out there”, cause them to intro-
duce security vulnerabilities in their code.
This seems to remain valid in our case: the se-
curity of some ransomware depends, at least in part,
on the security reliability of the unofficial sources. A
question remains open. Has the code-use helped ran-
somware criminals? The question is intertwined with
the practice of dual-use of research in the field and,
for this reason, we looked into the recent history of
ransomware attacks in search for episodes of code re-
use.
Dual-use Research in Ransomware Attacks: A Discussion on Ransomware Defence Intelligence
587
public string CreatePassword(int length)
{const string valid =
"a..zA..Z1234567890*!=&?&/";,
StringBuilder res = new StringBuilder();
Random rnd = new Random();
while (0 < length--)
res.Append(valid[rnd.Next(valid.Length)]);
return res.ToString();}
Figure 1: Password generation method of HiddenTear.
This password will later be used as the master secret to de-
rive encryption keys.
3 DUAL-USE & RANSOMWARE
Article 2 of Council Regulation (EC) No 428/2009
defines ‘dual-use items’ as items which can be used
for both civil and military purposes. The article in-
cludes “Computers” and “Telecommunications and
Information security” as categories to be screened for
potential dual-use.
When it comes to cryptography, dual-use is a se-
rious matter. In response to the US Munitions List,
Category XIII, Materials and Miscellaneous Articles,
which mentions “cryptographic devices, software and
components”, in a T-shirt shown at a DEFCON con-
ference it was reported provocatively a piece of (en-
cryption) code with the comment “this [code] can also
be a munition” (Herr and Rosenzweig, 2015).
Within the cryptography community there is
awareness that dual-use comes with a moral burden.
Rogaway wrote that “cryptography is an inherently
political tool, and it confers on the field an intrinsi-
cally moral dimension” (Rogaway, 2016). Rogaway’s
argument is scoped in the contention between privacy
on one side and mass surveillance on the other, but
the message on that DEFCON T-shirt extends, even
reverses, the matter. It raises the stake by pointing out
that cryptographic code can be misused as a weapon.
This is still the vision in certain countries, for instance
the US, where non-military cryptography exports are
if not forbidden at least controlled.
Being the subject of this paper ‘ransomware’, the
matter must be contextualized: what about dual-use
for cryptographic ransomware? And are ransomware
and their cryptographic components weapons? To
answer this question we look into cases of dual-use
in ransomware. The most controversial is that of
HiddenTear and its clones.
HiddenTear and Its Clones. In 2015, a Turk-
ish programmer Utku S¸en published the first fully-
private static string RandomString(int
length),
{string chars = "a..zA..Z0123456789";
StringBuilder stringBuilder = new
StringBuilder();,
Random random = new Random();
while (0 < length--)
stringBuilder.
Append(chars[random.Next(chars.Length)]);
return stringBuilder.ToString();}
Figure 2: Password generation method used by Tiggre.
The set of valid characters is shortened, most probably, to
ease the typing of the password when asked for recovery.
fledged, open-source ransomware HiddenTear. This
is the sample we commented in the previous section
and whose code to generate a password is shown in
Fig. 1.
From the early days, the release of HiddenTear
prototype received criticisms from the security com-
munity (Kovacs, 2016). The main concern of the re-
searchers is that even novice programmers can also
make use of the published ransomware code while
developing new variants. Time showed that they
were right. A McAfee researcher stated that “in
June (2017) almost 30% of the ‘new’ ransomware
species we discovered was based on the HiddenTear
code” (bee, ).
Three months after the first release, S¸en claimed
that he wished (i) to provide an example of ran-
somware for beginners (ii) to build a honeypot for
script kiddies (S¸en, 2015a). It was partly true that the
first variants of HiddenTear contained the same crit-
ical bugs that enabled the recovery of files (van der
Wiel, 2016). However, one real thing in the mal-
ware history is evolution. The bugs in the original
HiddenTear was fixed, and HiddenTear variant re-
placed the cryptographically insecure key generation
method with a new one (Trend Micro Blog, 2017)
which evades the state-of-the-art key-oriented anti-
ransomware defenses. Later, S¸en admitted that his
experiment was a total failure.
Another criticism to publishing the full source
codes of a ransomware regards the principle of re-
sponsible disclosure. Prior to sharing the sources, S¸en
did not informed the anti-virus vendors. It should be
noted that, when HiddenTear was released, on Au-
gust 2015, only a few anti-ransomware systems ex-
isted: signature-based detection was the main tech-
nique to stop ransomware, just as the other malware
types. Since HiddenTear and its variants were previ-
ously unseen, they were not recognized by AVs and
therefore could run undetected for a while. The only
ICISSP 2020 - 6th International Conference on Information Systems Security and Privacy
588
private void EncryptFile(string inputFile,
string outputFile),
try
{
// Your Key Here
string password = @"myKey123";
UnicodeEncoding UE = new
UnicodeEncoding();,
byte[] key = UE.GetBytes(password);
string cryptFile = outputFile;
FileStream fsCrypt = new
FileStream(cryptFile,
FileMode.Create);
,
,
RijndaelManaged RMCrypto = new
RijndaelManaged();,
CryptoStream cs = new
CryptoStream(fsCrypt,
RMCrypto.CreateEncryptor(key, key),
CryptoStreamMode.Write);
,
,
,
FileStream fsIn = new
FileStream(inputFile,
FileMode.Open);
,
,
int data;
while ((data = fsIn.ReadByte()) != -1)
cs.WriteByte((byte)data);
fsIn.Close();
cs.Close();
fsCrypt.Close();
}
catch
{
MessageBox.Show("Encryption failed!",
"Error");,
}}
Figure 3: A simple function to encrypt files with a pass-
word, published at CodeProject. Contrary to the common
practices, e.g., PBKDF2 (Kaliski, 2000), encryption key is
derived directly using UTF-16 character encoding. In addi-
tion, instead of generating a unique value, encryption key is
used as IV.
precaution S¸en took was putting a warning message
in HiddenTear source code, which cyber-criminals
could easily ignore.
Further Public Prototypes. S¸en is not the only
person that published a full ransomware prototype.
There are several ransomware projects in different
programming languages, publicly available on the
Internet. For instance, Arescrypt is another open
source ransomware implemented in C# (Fox, 2017).
GonnaCry is a Linux ransomware, implemented in
both C and Python (Marinho, 2017). Aiming at web
servers, a ransomware script written in PHP is also
available at (
ˇ
Sincek, 2019). There is even an “aca-
demic” ransomware prototype implemented in Go
language (de Souza Nunes, 2016). All these projects
are publicly available at GitHub, a well-known plat-
private static string GetEncKey()
{try
{
using (WebClient webClient = new
WebClient()),
return webClient.DownloadString(
@"http://ohad.000webhostapp.
com/cnc.php?txt=saveme")
.Trim();
,
,
,
}
catch
{
return "myke123!";
}}
private static void EncryptFile(string
inputFile, string outputFile, string
password)
,
,
{try
{
byte[] bytes = new UnicodeEncoding()
.GetBytes(password);,
FileStream fileStream1 = new
FileStream(outputFile,
FileMode.Create);
,
,
RijndaelManaged rijndaelManaged = new
RijndaelManaged();,
// [...]
fileStream1.Close();
System.IO.File.Delete(inputFile);
}
catch
{
Console.WriteLine("Error: Encryption
failed!");,
}}
Figure 4: File encryption function of the Crypren sample. If
C&C server is not reachable, which is shut down at the time
of this writing, the embedded password is used to derive
keys. The resemblances between hard-coded passwords,
key derivation methods and error messages are remarkable.
form among software developers. Moreover, although
S¸en abandoned the HiddenTear project, there are
still several clones of the original repository and even
some improved versions of HiddenTear on GitHub
website, for example (Rosa, 2017).
Zaitsev followed a different strategy when pub-
lishing CryptoTrooper (Zaitsev, 2016). He shared
the core part of the prototype as a closed source bi-
nary. The encryption algorithm, whose code was
not shared, contained a cryptographic flaw which en-
abled the recovery. Being closed source, the flaw
in the encryption module of CryptoTrooper could not
be fixed by the script-kiddies. Still, the commu-
nity was divided: some found the idea useful, oth-
ers did not (Cimpanu, 2016). In the end, Zaitsev re-
moved the project from GitHub but, as in the case of
HiddenTear, CryptoTrooper was forked by other de-
velopers. It is still accessible via various repositories.
Dual-use Research in Ransomware Attacks: A Discussion on Ransomware Defence Intelligence
589
All the developers of the publicly available ran-
somware prototypes states that their main motivation
was educational. However, a well documented ran-
somware code would also help to-be-cyber-criminals
to enter the ransomware business. Since ransomware
prototypes remain available on the Internet, the ethi-
cal question here is whether security researchers need
to publish and share full ransomware codes without
feeling accountable of the consequences, a recognized
ethical issue.
4 RANSOMWARE
INTELLIGENCE
Herr and Rosenzweig suggest that a piece of code
is cyber-weapon when it combines “propagation, ex-
ploitation, and payload [i.e., damaging] capabilities”
(Herr and Rosenzweig, 2015). Each components,
despite innocuous in separation, carry the potential-
ity to be combined with the missing others into a
weapon. However, to have a military use, a software
must create or tangibly support the deployment of
destructive effects. These could be short term, where
deleted data is restored from backup, or near perma-
nent, where a payload is designed to damage a de-
vice’s firmware” (Herr and Rosenzweig, 2015).
Ransomware may have such a destructive effect.
For sake of an example, at the time of the writing,
June 2019, the major electricity supplier in South
Africa’s city of Johannesburg was attached, leaving
more than a quarter of a million people in the dark.
Another attack forced a shutdown of its websites and
billing systems as a precautionary measure.
Ransomware variants, called wipeware, can wipe
data clean. Allegedly deployed to attack Saudi en-
ergy companies and Iranian oil companies, they had
destructive consequences. One variant of it, Shamoon
wiper, has been released to attack Sony Pictures En-
tertainment, succeeding to avoid the outing of ‘The
Interview’, a documentary mocking the North Ko-
rean dictator, Kim Jong-un. If we adhere to Schmid’s
claims that “terrorist violence is predominantly polit-
ical” (Schmid, 2011), such events can be considered
also “terrorist attack” .
If ransomware are to be regarded as cyber-
weapons, as we claim, could it be conceivable to ap-
ply intelligence and counter-intelligence strategies to
mitigate the threats and control the consequences of
an attack? And, if yes, how?
Cyber-Intelligence has been defined as “the pro-
cess by which specific types of information important
to national security are requested, collected, analyzed,
and provided to policymakers, the products of that
process”(Lowenthal, 2016). Duvenage et al. (Duve-
nage et al., 2015), call this positive intelligence, to
distinguish it from counter-intelligence, which is the
countering of an hostile intelligence activity.
Ransomware Positive Intelligence. For ran-
somware threat, positive intelligence could consist in
gathering information about modalities of working.
It should be about how the ransomware propagates,
exploits vulnerabilities, and executes it payload. In
the Open Source Intelligence (OSINT), several initia-
tives exist aiming to collect and analyse information
gathered from public or open sources. An example
is the NoMoreRansom project
1
. It aims to inform the
public and to collect incidents reports, including to
gather the information from public platforms that can
be potentially utilized by ransomware authors. Other
platforms, although not specifically dedicated to ran-
somware, such as the Malware Information Sharing
Platform (MISP)
2
—a free and open source software
helping information sharing of threat intelligence,
including cyber-security indicators—can offer tools
that enable intelligence analysis. Such platforms can
be employed to control the information flow during
an attack, spreading alerts following a Warning and
Coordination action, and to help potential victims
“raise their shields” as soon as possible.
Ransomware Counter-intelligence. According
to (Coleman, 2009), Counter Cyber Intelligence
(CCI) is the ensemble of “all efforts made by one
intelligence organization to prevent adversaries,
enemy intelligence organizations or criminal orga-
nizations from gathering and collecting sensitive
digital information or intelligence about them via
computers, networks and associated equipment”. It
can be implemented using strategies that, according
to Panda Security, a cyber-security company, either
consists of “leaving doors open” (i.e., left access
points unprotected on purpose), “inject fake infor-
mation” (i.e., fake confidential information), and
“keeping them busy while stealing” (i.e., watching
and obtaining information about the attacker).
Looking into the internet and searching for
“counter-intelligence for ransomware”, we have
found that the majority of the initiatives to pro-
tect from ransomware attacks focuses on rais-
ing awareness. For instance, the US National
Counter-intelligence and Security Center (NCSC) has
launched in January 2019 a campaign “Know the
Risk, Raise Your Shield”. The Cybersecurity and In-
1
https://www.nomoreransom.org/en/
2
https://www.misp-project.org/.
ICISSP 2020 - 6th International Conference on Information Systems Security and Privacy
590
frastructure Security Agency (CISA) addresses ran-
somware specifically, but it is all about knowing
the threat and apply general security best practices
such as backing-up data. We have found, within
the scope of cryptrographic ransomware and limit-
edly to this on-going work, nothing about “leaving
the door open”, “inject fake information”, or ‘’keep
them busy”.
The second measure (i.e., “inject fake informa-
tion”) may not be fully applicable at least if that
means to avoid to spread knowledge about how to
build the ransomware weapon: the instruments of
cryptography are nowadays already known and pub-
lic. However, at the light of what we have discussed
in the previous sections, it may be a strategy to post
the code of variants whose decryptors already exist.
For what concerns the “keep them busy”
paradigm, as discussed in (Genc¸ et al., 2019), it may
be possible the use decoy files to deflect a ransomware
attack against irrelevant (for the victim) files, so gain-
ing that amount of time required to stop the attack’s
development. Using decoy files could be paired
with strategies that downgrade the efficiency of en-
cryption for applications that are not trustworthy or
whitelisted. We have not investigated in this direc-
tion, but this option seems preferable to that of run-
ning untrusted application in sandbox. This can be
less effective, since certain ransomware sample rec-
ognize the presence of a virtual environment and re-
main dormant. A few articles suggest the use of Arti-
ficial Intelligence (e.g., (Huang et al., 2018)), but we
did not look into this direction.
5 CONCLUSION
Ransomware are emerging as cyber-weapons. They
have been used in attacks that resemble actions of
cyber-war, and are far more dangerous and disruptive
than traditional malware. Consequently, the research
community should reflect on coordinated actions to
address the threat under an appropriate code of ethi-
cal conduct.
Having discovered that a few ransomware con-
tain a copy-paste from cryptographic code available
in public sources, we debated the matter of dual-use
in cryptographic research and recalled (in)famous an-
tecedents in the recent ransomware history. Since we
managed to build decryptors for those ransomware,
the dual use turned out to be a double-edge for the
criminals, but generally it is not. After having build a
case for ransomware as cyber-weapon, we briefly re-
viewed intelligence and counter-intelligent strategies
that could be used in the fight against ransomware.
We did not backed our speculations with field
studies or interviews. Ours is an educated argumenta-
tion, but its purpose is to invite the anti-ransomware
community to be more proactive in the cyberwar
against ransomware. Even the excellent NoMoreRan-
som project, which offers decryptors when they are
available (as did in June 2019, with the latest version
of Gandcrab
3
), at the end of the day praises for keep-
ing back-up, within a “Better Safe Than Sorry” ad-
vice.
REFERENCES
Acar, Y., Backes, M., Fahl, S., Kim, D., Mazurek, M. L.,
and Stransky, C. (2017). How Internet Resources
Might Be Helping You Develop Faster but Less Se-
curely. IEEE Security and Privacy, 15(2):50–60.
Cimpanu, C. (2016). New Open Source Linux
Ransomware Shows Infosec Community
Divide. https://news.softpedia.com/news/
new-open-source-linux-ransomware-shows\
-infosec-community-divide-508669.shtml. [20-
Jul-2019].
Coleman, K. (2009). Counter Cyber Intelligence.
https://www.military.com/defensetech/2009/03/09/
counter-cyber-intelligence. [20-Jul-2019].
S¸en, U. (2015a). Destroying The Encryption of Hidden Tear
Ransomware. https://utkusen.com/blog/destroying\
-the-encryption-of-hidden-tear-ransomware.html.
[21-Jul-201].
S¸en, U. (2015b). Hiddentear: an open source ransomware-
like file crypter kit. https://github.com/utkusen/
hidden-tear.
de Souza Nunes, M. (2016). Ransomware: A POC Win-
dows crypto-ransomware (Academic). https://github.
com/mauri870/ransomware. [20-Jul-2019].
den Bosch, T. V. (2005). Encrypt/Decrypt Files in
VB.NET (Using Rijndael). https://www.codeproject.
com/Articles/12092/Encrypt-Decrypt-Files-in-VB\
-NET-Using-Rijndael. [20-Jul-2019].
dnSpy (2019). .NET debugger and assembly editor. https:
//github.com/0xd4d/dnSpy. [2019].
Duvenage, P., von Solms, S., and Corregedor, M. (2015).
The Cyber Counterintelligence Process: A Concep-
tual Overview and Theoretical Proposition. In Proc.
of the 14th ECCWS, pages 42–52. ACPI.
Fischer, F., B
¨
ottinger, K., Xiao, H., Stransky, C., Acar, Y.,
Backes, M., and Fahl, S. (2017). Stack Overflow Con-
sidered Harmful? The Impact of Copy Paste on An-
droid Application Security. In 2017 IEEE SP, pages
121–136.
Fox, W. (2017). Arescrypt: Experimental ransomware for
windows 7+ with aes-256 support. https://github.com/
BlackVikingPro/arescrypt. [20-Jul-2019].
3
https://www.malwarebytes.com/gandcrab/
Dual-use Research in Ransomware Attacks: A Discussion on Ransomware Defence Intelligence
591
Genc¸, Z. A., Lenzini, G., and Sgandurra, D. (2019).
On deception-based protection against cryptographic
ransomware. In Detection of Intrusions and Mal-
ware, and Vulnerability Assessment, pages 219–239.
Springer.
Herr, T. and Rosenzweig, P. (2015). Cyber Weapons & Ex-
port Control: Incorporating Dual Use with the PrEP
Model. J. National Security Law and Policy, 8(2).
Huang, D. Y., Aliapoulios, M. M., Li, V. G., Invernizzi, L.,
Bursztein, E., McRoberts, K., Levin, J., Levchenko,
K., Snoeren, A. C., and McCoy, D. (2018). Tracking
ransomware end-to-end. In IEEE Security and Pri-
vacy, pages 618–631. IEEE.
Hybrid Analysis (2019). Free Automated Malware Analysis
Service. https://www.hybrid-analysis.com/. [2019].
Johnson, M. (2008). How do you do Impersonation in
.NET? (rev. 2). https://stackoverflow.com/revisions/
7250145/2. [20-Jul-2019].
Kaliski, B. (2000). PKCS #5: Password-Based Cryptogra-
phy Specification Version 2.0. RFC 2898.
Kovacs, E. (2016). Educational Ransomware Abused by
Cybercriminals. https://www.securityweek.com\
/educational\-ransomware-abused-cybercriminals.
[20-Jul-2019].
Lowenthal, M. M. (2016). Intelligence: From Secrets to
Policy. CQ Press, Los Angeles, 7 edition.
Lydford, S. (2008). File Encryption and Decryp-
tion in C#. https://www.codeproject.com/Articles/
26085/File-Encryption-and-Decryption-in-C. [20-
Jul-2019].
Marinho, T. (2017). GonnaCry: A Linux Ransomware.
https://github.com/tarcisio-marinho/GonnaCry. [20-
Jul-2019].
Rogaway, P. (2016). The Moral Character of Cryptographic
Work. Austin, TX. USENIX Association.
Rosa, A. (2017). Hiddentear (forked). https://github.com/
Virgula0/hidden-tear. [20-Jul-2019].
Schmid, A. P. (2011). The Definition of Terrorism. In The
Routledge Handbook of Terrorism Research, chap-
ter 2, pages 39–157. Routledge, Oxon, UK.
Trend Micro Blog (2017). Ransomware Recap: The
Ongoing Development of Hidden Tear Vari-
ants. https://www.trendmicro.com/vinfo/us/
security/news/cybercrime-and-digital-threats/
the-ongoing-development-of-hidden\-tear-variants.
[20-Jul-2019].
van der Wiel, J. (2016). Hidden tear and its spin offs.
https://securelist.com/hidden-tear-and-its-spin-offs/
73565/. [20-Jul-2019].
VirusTotal Threat Intelligence (2019). Virustotal. https:
//www.virustotal.com/.
ˇ
Sincek, I. (2019). Ransomware: PHP ransomware that en-
crypts your files as well as file and directory names.
https://github.com/ivan-sincek/ransomware. [20-Jul-
2019].
Zaitsev, M. (2016). CryptoTrooper: The world’s first Linux
white-box ransomware. https://github.com/cryptolok/
CryptoTrooper. [20-Jul-2019].
ICISSP 2020 - 6th International Conference on Information Systems Security and Privacy
592