Void	O
Rabisu	O
’s	O
Use	O
of	O
RomCom	O
Backdoor	O
Shows	O
a	O

	O
Growing	O
Shift	O
in	O
Threat	O
Actors	O
’	O
Goals	O

	O
trendmicro.com/en_us/research/23/e/void-rabisu-s-use-of-romcom-backdoor-shows-a-growing-shift-in-th.html	O

	O
May	O
30	O
,	O
2023	O


	O
Cyber	O
Threats	O

	O
Void	O
Rabisu	O
,	O
a	O
malicious	O
actor	O
believed	O
to	O
be	O
associated	O
with	O
the	O
RomCom	O
backdoor	O
,	O
was	O

	O
thought	O
to	O
be	O
driven	O
by	O
financial	O
gain	O
because	O
of	O
its	O
ransomware	O
attacks	O
.	O
But	O
in	O
this	O
blog	O

	O
entry	O
,	O
we	O
discuss	O
how	O
the	O
use	O
of	O
the	O
RomCom	O
backdoor	O
in	O
recent	O
attacks	O
shows	O
how	O
Void	O

	O
Rabisu	O
's	O
motives	O
seem	O
to	O
have	O
changed	O
since	O
at	O
least	O
October	O
2022	O
.	O

	O
By	O
:	O
Feike	O
Hacquebord	O
,	O
Stephen	O
Hilt	O
,	O
Fernando	O
Merces	O
,	O
Lord	O
Alfred	O
Remorin	O
May	O
30	O
,	O
2023	O

	O
Read	O
time	O
:	O
(	O
words	O
)	O


	O
With	O
contributions	O
from	O
Veronica	O
Chierzi	O
and	O
Jayvee	O
Mark	O
Villaroman	O

	O
Since	O
the	O
start	O
of	O
the	O
war	O
in	O
Ukraine	O
in	O
February	O
2022	O
,	O
the	O
number	O
of	O
cyber	O
campaigns	O

	O
against	O
Ukraine	O
and	O
North	O
Atlantic	O
Treaty	O
Organization	O
(	O
NATO	O
)	O
countries	O
has	O
increased	O

	O
significantly	O
.	O
These	O
campaigns	O
come	O
from	O
many	O
different	O
angles	O
:	O
known	O
advanced	O
persistent	O

	O
threat	O
(	O
APT	O
)	O
actors	O
,	O
APT	O
actors	O
that	O
were	O
not	O
publicly	O
reported	O
on	O
before	O
,	O
and	O
cyber	O

	O
mercenaries	O
,	O
hacktivists	O
,	O
and	O
criminal	O
actors	O
who	O
appear	O
to	O
have	O
shifted	O
from	O
purely	O
financial	O

	O
motives	O
to	O
geopolitical	O
goals	O
.	O
In	O
the	O
past	O
,	O
these	O
actors	O
had	O
different	O
motivations	O
,	O
mode	O
of	O

	O
operations	O
,	O
and	O
targets	O
,	O
but	O
the	O
line	O
between	O
their	O
campaigns	O
has	O
started	O
to	O
blur	O
:	O
Not	O
only	O
is	O

	O
an	O
overlap	O
in	O
their	O
targeting	O
becoming	O
apparent	O
,	O
but	O
the	O
distinction	O
between	O
their	O
modes	O
of	O

	O
operation	O
is	O
less	O
clear	O
.	O
For	O
instance	O
,	O
in	O
2022	O
,	O
one	O
of	O
Conti	O
’s	O
affiliates	O
was	O
found	O
to	O
be	O
using	O

	O
its	O
initial	O
access	O
techniques	O
against	O
Ukraine	O
instead	O
of	O
using	O
them	O
to	O
spread	O
ransomware	O
.	O

	O
Another	O
example	O
of	O
this	O
is	O
Void	O
Rabisu	O
,	O
also	O
known	O
as	O
Tropical	O
Scorpius	O
,	O
an	O
actor	O
believed	O

	O
to	O
be	O
associated	O
with	O
Cuba	O
ransomware	O
and	O
the	O
RomCom	O
backdoor	O
.	O
Because	O
of	O
its	O
many	O

	O
ransomware	O
attacks	O
,	O
Void	O
Rabisu	O
was	O
believed	O
to	O
be	O
financially	O
motivated	O
,	O
even	O
though	O
its	O

	O
associated	O
Cuba	O
ransomware	O
allegedly	O
attacked	O
the	O
parliament	O
of	O
Montenegro	O
in	O
August	O

	O
2022	O
,	O
which	O
could	O
be	O
considered	O
part	O
of	O
a	O
geopolitical	O
agenda	O
.	O
The	O
motives	O
of	O
Void	O
Rabisu	O

	O
seem	O
to	O
have	O
changed	O
since	O
at	O
least	O
October	O
2022	O
,	O
when	O
Void	O
Rabisu	O
’s	O
associated	O

	O
RomCom	O
backdoor	O
was	O
reported	O
to	O
have	O
been	O
used	O
in	O
attacks	O
against	O
the	O
Ukrainian	O

	O
government	O
and	O
military	O
:	O
In	O
a	O
campaign	O
in	O
December	O
2022	O
,	O
a	O
fake	O
version	O
of	O
the	O
Ukrainian	O

	O
army	O
’s	O
DELTA	O
situational	O
awareness	O
website	O
was	O
used	O
to	O
lure	O
targets	O
into	O
installing	O
the	O

	O
RomCom	O
backdoor	O
.	O
Normally	O
,	O
this	O
kind	O
of	O
brazen	O
attack	O
would	O
be	O
thought	O
to	O
be	O
the	O
work	O
of	O
a	O


	O
1/27	O


	O
nation	O
state	O
-	O
sponsored	O
actor	O
,	O
but	O
in	O
this	O
case	O
,	O
the	O
indicators	O
clearly	O
pointed	O
towards	O
Void	O

	O
Rabisu	O
,	O
and	O
some	O
of	O
the	O
tactics	O
,	O
techniques	O
,	O
and	O
procedures	O
(	O
TTPs	O
)	O
used	O
were	O
typically	O

	O
associated	O
with	O
cybercrime	O
.	O

	O
Trend	O
Micro	O
’s	O
telemetry	O
and	O
research	O
corroborates	O
that	O
the	O
RomCom	O
backdoor	O
has	O
been	O

	O
used	O
in	O
geopolitically	O
motivated	O
attacks	O
since	O
at	O
least	O
October	O
2022	O
,	O
with	O
targets	O
that	O

	O
included	O
organizations	O
in	O
Ukraine	O
’s	O
energy	O
and	O
water	O
utility	O
sectors	O
.	O
Targets	O
outside	O
of	O

	O
Ukraine	O
were	O
observed	O
as	O
well	O
,	O
such	O
as	O
a	O
provincial	O
local	O
government	O
that	O
provides	O
help	O
to	O

	O
Ukrainian	O
refugees	O
,	O
a	O
parliament	O
member	O
of	O
a	O
European	O
country	O
,	O
a	O
European	O
defense	O

	O
company	O
,	O
and	O
various	O
IT	O
service	O
providers	O
in	O
Europe	O
and	O
the	O
US	O
.	O
Independent	O
research	O
from	O

	O
Google	O
showed	O
that	O
RomCom	O
was	O
being	O
used	O
in	O
campaigns	O
against	O
attendees	O
of	O
the	O

	O
Masters	O
of	O
Digital	O
conference	O
,	O
a	O
conference	O
organized	O
by	O
DIGITALEUROPE	O
,	O
and	O
the	O
Munich	O

	O
Security	O
Conference	O
.	O

	O
In	O
this	O
blog	O
entry	O
,	O
we	O
will	O
discuss	O
how	O
the	O
use	O
of	O
the	O
RomCom	O
backdoor	O
fits	O
into	O
the	O
current	O

	O
landscape	O
,	O
where	O
politically	O
motivated	O
attacks	O
are	O
not	O
committed	O
by	O
nation	O
-	O
state	O
actors	O

	O
alone	O
.	O
Even	O
though	O
we	O
can	O
not	O
confirm	O
coordination	O
between	O
the	O
different	O
attacks	O
,	O
Ukraine	O

	O
and	O
countries	O
who	O
support	O
Ukraine	O
are	O
being	O
targeted	O
by	O
various	O
actors	O
,	O
like	O
APT	O
actors	O
,	O

	O
hacktivists	O
,	O
cyber	O
mercenaries	O
,	O
and	O
cybercriminals	O
like	O
Void	O
Rabisu	O
.	O
We	O
will	O
also	O
delve	O
into	O

	O
how	O
RomCom	O
has	O
evolved	O
over	O
time	O
and	O
how	O
the	O
backdoor	O
is	O
spread	O
both	O
by	O
methods	O
that	O

	O
look	O
like	O
APT	O
,	O
as	O
well	O
as	O
methods	O
used	O
by	O
prominent	O
cybercriminal	O
campaigns	O
taking	O
place	O

	O
currently	O
,	O
to	O
show	O
that	O
RomCom	O
is	O
using	O
more	O
detection	O
evasion	O
techniques	O
that	O
are	O
popular	O

	O
among	O
the	O
most	O
impactful	O
cybercriminals	O
.	O

	O
We	O
assess	O
that	O
RomCom	O
makes	O
use	O
of	O
the	O
same	O
third	O
-	O
party	O
services	O
that	O
are	O
being	O
utilized	O

	O
by	O
other	O
criminal	O
actors	O
as	O
well	O
,	O
like	O
malware	O
signing	O
and	O
binary	O
encryption	O
.	O
RomCom	O
has	O

	O
been	O
spread	O
through	O
numerous	O
lure	O
sites	O
that	O
are	O
sometimes	O
set	O
up	O
in	O
rapid	O
bursts	O
.	O
These	O

	O
lure	O
sites	O
are	O
most	O
likely	O
only	O
meant	O
for	O
a	O
small	O
number	O
of	O
targets	O
,	O
thus	O
making	O
discovery	O

	O
and	O
analysis	O
more	O
difficult	O
.	O
Void	O
Rabisu	O
is	O
one	O
of	O
the	O
most	O
evident	O
examples	O
of	O
financially	O

	O
motivated	O
threat	O
actors	O
whose	O
goals	O
and	O
motivations	O
are	O
becoming	O
more	O
aligned	O
under	O

	O
extraordinary	O
geopolitical	O
circumstances	O
,	O
and	O
we	O
anticipate	O
that	O
this	O
will	O
happen	O
more	O
in	O
the	O

	O
future	O
.	O

	O
RomCom	O
campaigns	O

	O
We	O
have	O
been	O
tracking	O
RomCom	O
campaigns	O
since	O
the	O
summer	O
of	O
2022	O
,	O
and	O
since	O
then	O
,	O

	O
have	O
seen	O
an	O
escalation	O
in	O
its	O
detection	O
evasion	O
methods	O
:	O
Not	O
only	O
do	O
the	O
malware	O
samples	O

	O
routinely	O
use	O
VMProtect	O
to	O
make	O
both	O
manual	O
and	O
automated	O
sandbox	O
analysis	O
more	O

	O
difficult	O
,	O
they	O
also	O
utilize	O
binary	O
padding	O
techniques	O
on	O
the	O
payload	O
files	O
.	O
This	O
adds	O
a	O

	O
significant	O
amount	O
of	O
overlay	O
bytes	O
to	O
the	O
files	O
,	O
increasing	O
the	O
size	O
of	O
the	O
malicious	O
payload	O

	O
(	O
we	O
've	O
seen	O
a	O
file	O
with	O
1.7	O
gigabytes	O
)	O
.	O
Additionally	O
,	O
a	O
new	O
routine	O
has	O
been	O
recently	O
added	O

	O
that	O
involves	O
the	O
encryption	O
of	O
the	O
payload	O
files	O
,	O
which	O
can	O
only	O
be	O
decrypted	O
if	O
a	O
certain	O
key	O

	O
is	O
downloaded	O
to	O
activate	O
the	O
payload	O
.	O


	O
2/27	O


	O
In	O
addition	O
to	O
these	O
technical	O
evasion	O
techniques	O
,	O
RomCom	O
is	O
being	O
distributed	O
using	O
lure	O

	O
sites	O
that	O
often	O
appear	O
legitimate	O
and	O
are	O
being	O
utilized	O
in	O
narrow	O
targeting	O
.	O
This	O
makes	O

	O
automated	O
blocking	O
of	O
these	O
lure	O
websites	O
through	O
web	O
reputation	O
systems	O
harder	O
.	O
Void	O

	O
Rabisu	O
has	O
been	O
using	O
Google	O
Ads	O
to	O
entice	O
their	O
targets	O
to	O
visit	O
the	O
lure	O
sites	O
,	O
similar	O
to	O
a	O

	O
campaign	O
that	O
distributed	O
IcedID	O
botnet	O
in	O
December	O
2022	O
.	O
A	O
key	O
difference	O
is	O
that	O
while	O

	O
IcedID	O
’s	O
targeting	O
was	O
wider	O
,	O
Void	O
Rabisu	O
probably	O
opted	O
for	O
narrower	O
targeting	O
that	O
Google	O

	O
Ads	O
offers	O
to	O
its	O
advertisers	O
.	O
RomCom	O
campaigns	O
also	O
make	O
use	O
of	O
highly	O
targeted	O
spear	O

	O
phishing	O
emails	O
.	O

	O
On	O
the	O
RomCom	O
lure	O
sites	O
,	O
targets	O
are	O
offered	O
trojanized	O
versions	O
of	O
legitimate	O
applications	O
,	O

	O
like	O
chat	O
apps	O
such	O
as	O
AstraChat	O
and	O
Signal	O
,	O
PDF	O
readers	O
,	O
remote	O
desktop	O
apps	O
,	O
password	O

	O
managers	O
,	O
and	O
other	O
tools	O
,	O
that	O
are	O
typically	O
used	O
by	O
system	O
administrators	O
.	O

	O
dirwinstat.com	O

	O
(	O
as	O
of	O
April	O
4	O
,	O
2023	O
)	O


	O
3/27	O


	O
devolrdm.com	O

	O
(	O
as	O
of	O
March	O
23	O
,	O
2023	O
)	O


	O
vectordmanagesoft.com	O

	O
(	O
as	O
of	O
March	O
22	O
,	O
2023	O
)	O


	O
4/27	O


	O
cozy-sofware.com	O

	O
(	O
as	O
of	O
March	O
13	O
,	O
2023	O
)	O


	O
devolutionrdp.com	O

	O
(	O
as	O
of	O
March	O
6	O
,	O
2023	O
)	O


	O
5/27	O


	O
astrachats.com	O

	O
(	O
as	O
of	O
February	O
27	O
,	O

	O
2023	O
)	O


	O
chatgpt4beta.com	O

	O
(	O
as	O
of	O
February	O
23	O
,	O

	O
2023	O
)	O


	O
6/27	O


	O
singularlabs.org	O

	O
(	O
as	O
of	O
January	O
30	O
,	O

	O
2023	O
)	O


	O
gotomeet.us	O

	O
(	O
as	O
of	O
December	O
14	O
,	O

	O
2022	O
)	O


	O
7/27	O


	O
gllmp.com	O

	O
(	O
as	O
of	O
December	O
8	O
,	O

	O
2022	O
)	O


	O
lnfo-messengers.com	O

	O
(	O
as	O
of	O
November	O
3	O
,	O

	O
2022	O
)	O


	O
8/27	O


	O
pass-shield.com	O

	O
(	O
as	O
of	O
October	O
15	O
,	O

	O
2022	O
)	O


	O
pdffreader.com	O

	O
(	O
as	O
of	O
October	O
12	O
,	O

	O
2022	O
)	O


	O
9/27	O


	O
10/27	O


	O
veeame.com	O

	O
(	O
as	O
of	O
September	O
,	O
9	O

	O
2022	O
)	O


	O
npm-solar.com	O

	O
(	O
as	O
of	O
July	O
31	O
,	O
2022	O
)	O


	O
11/27	O


	O
advanced-ipscanners.com	O

	O
(	O
as	O
of	O
July	O
20	O
,	O
2022	O
)	O


	O
Table	O
1	O
.	O
RomCom	O
lure	O
sites	O

	O
Image	O
credit	O
:	O
DomainTools	O

	O
As	O
reported	O
by	O
the	O
Ukrainian	O
Computer	O
Emergency	O
Response	O
Team	O
(	O
CERT	O
-	O
UA	O
)	O
in	O
the	O
fall	O
of	O

	O
2022	O
,	O
RomCom	O
was	O
used	O
in	O
specific	O
campaigns	O
against	O
Ukrainian	O
targets	O
,	O
including	O
the	O

	O
Ukrainian	O
government	O
and	O
the	O
Ukrainian	O
military	O
.	O
Trend	O
Micro	O
’s	O
telemetry	O
confirms	O
this	O

	O
targeting	O
,	O
and	O
,	O
as	O
shown	O
in	O
a	O
selection	O
of	O
the	O
numerous	O
RomCom	O
campaigns	O
over	O
time	O

	O
(	O
Table	O
1	O
)	O
,	O
it	O
is	O
immediately	O
clear	O
that	O
RomCom	O
already	O
had	O
Ukrainian	O
-	O
language	O
social	O

	O
engineering	O
lures	O
back	O
in	O
October	O
and	O
November	O
2022	O
.	O

	O
We	O
count	O
a	O
few	O
dozen	O
lure	O
websites	O
that	O
have	O
been	O
set	O
up	O
since	O
July	O
2022	O
.	O
RomCom	O
shows	O

	O
a	O
mix	O
in	O
their	O
targeting	O
methodologies	O
,	O
mixing	O
typical	O
cybercriminal	O
TTPs	O
with	O
TTPs	O
that	O
are	O

	O
more	O
common	O
for	O
APT	O
actors	O
.	O
For	O
example	O
,	O
RomCom	O
used	O
spear	O
phishing	O
against	O
a	O

	O
member	O
of	O
a	O
European	O
parliament	O
in	O
March	O
2022	O
,	O
but	O
targeted	O
a	O
European	O
defense	O

	O
company	O
in	O
October	O
2022	O
with	O
a	O
Google	O
Ads	O
advertisement	O
that	O
led	O
to	O
an	O
intermediary	O

	O
landing	O
site	O
that	O
would	O
redirect	O
to	O
a	O
RomCom	O
lure	O
site	O
.	O
That	O
intermediary	O
landing	O
site	O
used	O

	O
the	O
domain	O
name	O
“	O
kagomadb[.]com	O
,	O
”	O
which	O
was	O
later	O
used	O
for	O
Qakbot	O
and	O
Gozi	O
payloads	O
in	O

	O
December	O
2022	O
.	O


	O
12/27	O


	O
Among	O
the	O
targets	O
we	O
have	O
seen	O
based	O
on	O
Trend	O
Micro	O
’s	O
telemetry	O
were	O
a	O
water	O
utility	O

	O
company	O
,	O
entities	O
in	O
the	O
financial	O
and	O
energy	O
sectors	O
,	O
and	O
an	O
IT	O
company	O
in	O
Ukraine	O
.	O
Outside	O

	O
Ukraine	O
,	O
other	O
targets	O
included	O
a	O
local	O
government	O
agency	O
that	O
supports	O
Ukrainian	O
refugees	O
,	O

	O
a	O
defense	O
company	O
in	O
Europe	O
,	O
a	O
high	O
-	O
profile	O
European	O
politician	O
,	O
several	O
IT	O
service	O

	O
providers	O
in	O
Europe	O
and	O
the	O
US	O
,	O
a	O
bank	O
in	O
South	O
America	O
,	O
and	O
a	O
couple	O
of	O
targets	O
located	O
in	O

	O
Asia	O
.	O
Combined	O
with	O
the	O
targets	O
that	O
were	O
published	O
by	O
CERT	O
-	O
UA	O
and	O
Google	O
,	O
a	O
clear	O

	O
picture	O
emerges	O
of	O
the	O
RomCom	O
backdoor	O
’s	O
targets	O
:	O
select	O
Ukrainian	O
targets	O
and	O
allies	O
of	O

	O
Ukraine	O
.	O

	O
RomCom	O
3.0	O
:	O
The	O
AstraChat	O
Campaign	O

	O
In	O
this	O
section	O
,	O
we	O
will	O
analyze	O
one	O
of	O
the	O
RomCom	O
backdoor	O
samples	O
that	O
was	O
used	O
in	O

	O
February	O
2023	O
against	O
targets	O
in	O
Eastern	O
Europe	O
.	O
Previous	O
RomCom	O
versions	O
analyzed	O
by	O

	O
Palo	O
Alto	O
's	O
Unit	O
42	O
use	O
a	O
modular	O
architecture	O
and	O
support	O
up	O
to	O
20	O
different	O
commands	O
.	O

	O
Since	O
then	O
,	O
the	O
malware	O
evolved	O
significantly	O
in	O
terms	O
of	O
the	O
number	O
of	O
supported	O

	O
commands	O
,	O
but	O
its	O
modular	O
architecture	O
remains	O
almost	O
unchanged	O
.	O
The	O
threat	O
actor	O
behind	O

	O
RomCom	O
3.0	O
also	O
makes	O
use	O
of	O
different	O
techniques	O
to	O
drop	O
and	O
execute	O
the	O
malware	O
.	O
This	O

	O
analysis	O
is	O
based	O
on	O
a	O
campaign	O
that	O
embedded	O
RomCom	O
3.0	O
in	O
an	O
AstraChat	O
instant	O

	O
messaging	O
software	O
installation	O
package	O
.	O

	O
Dropper	O

	O
The	O
file	O
astrachat.msi	O
is	O
a	O
Microsoft	O
Installer	O
(	O
MSI	O
)	O
archive	O
.	O
Despite	O
installing	O
files	O
related	O
to	O

	O
legitimate	O
AstraChat	O
software	O
,	O
it	O
unpacks	O
a	O
malicious	O
InstallA.dll	O
file	O
and	O
calls	O
its	O
Main	O
(	O
)	O

	O
function	O
(	O
Figure	O
1	O
)	O
.	O


	O
Figure	O
1	O
.	O
CustomAction	O
table	O
from	O
a	O
RomCom	O
MSI	O
dropper	O

	O
The	O
InstallA.dll	O
file	O
extracts	O
three	O
Dynamic	O
Link	O
Libraries	O
(	O
DLLs	O
)	O
files	O
under	O
the	O

	O
%	O
PUBLIC%\Libraries	O
folder	O
:	O

	O
prxyms	O
<	O
number>.dll	O

	O
13/27	O


	O
winipfile	O
<	O
number>.dll0	O

	O
netid	O
<	O
number>.dll0	O

	O
The	O
number	O
in	O
these	O
DLL	O
files	O
is	O
an	O
integer	O
number	O
based	O
on	O
the	O
Machine	O
GUID	O
read	O
from	O

	O
Windows	O
Registry	O
at	O
HKLM\SOFTWARE\Microsoft\Cryptography\MachineGuid	O
.	O

	O
Persistence	O

	O
For	O
persistence	O
,	O
RomCom	O
uses	O
COM	O
hijacking	O
,	O
hence	O
its	O
name	O
.	O
InstallA.dll	O
writes	O
the	O

	O
following	O
registry	O
value	O
in	O
Windows	O
Registry	O
:	O

	O
[	O
HKEY_CURRENT_USER\SOFTWARE\Classes\CLSID\{C90250F3	O
-	O
4D7D-4991	O
-	O
9B69A5C5BC1C2AE6}\InProcServer32	O
]	O

	O
@="%PUBLIC%\\Libraries\\prxyms	O
<	O
number>.dll	O
"	O

	O
This	O
overwrites	O
the	O
same	O
key	O
under	O
the	O
HKEY_LOCAL_MACHINE	O
hive	O
,	O
causing	O
processes	O

	O
that	O
request	O
this	O
Class	O
ID	O
(	O
CLSID	O
)	O
to	O
load	O
the	O
RomCom	O
loader	O
DLL	O
at	O

	O
%	O
PUBLIC%\Libraries\prxyms	O
<	O
number>.dll	O
.	O
One	O
such	O
process	O
is	O
explorer.exe	O
,	O
which	O
is	O

	O
restarted	O
by	O
RomCom	O
dropper	O
,	O
so	O
the	O
loader	O
DLL	O
is	O
called	O
.	O

	O
The	O
RomCom	O
loader	O
also	O
redirects	O
calls	O
to	O
its	O
exported	O
functions	O
to	O
the	O
legit	O
actxprxy.dll	O
by	O

	O
making	O
use	O
of	O
forwarded	O
exports	O
(	O
Figure	O
2	O
)	O
.	O


	O
Figure	O
2	O
.	O
Forwarded	O
exports	O
from	O
the	O
RomCom	O
3.0	O
loader	O
(	O
prxyms	O
<	O
number>.dll	O
)	O

	O
However	O
,	O
before	O
a	O
call	O
is	O
forwarded	O
,	O
the	O
malicious	O
code	O
at	O
the	O
DLL	O
entry	O
point	O
of	O
RomCom	O

	O
loader	O
runs	O
.	O
This	O
code	O
uses	O
rundll32.exe	O
to	O
execute	O
the	O
exported	O
Main	O
(	O
)	O
function	O
from	O
both	O

	O
winipfile	O
<	O
number>.dll0	O
and	O
netid	O
<	O
number>.dll0	O
.	O

	O
Architecture	O

	O
RomCom	O
3.0	O
is	O
divided	O
into	O
three	O
components	O
:	O
a	O
loader	O
,	O
a	O
network	O
component	O
that	O
interacts	O

	O
with	O
the	O
command	O
-	O
and	O
-	O
control	O
(	O
C&C	O
)	O
server	O
,	O
and	O
a	O
worker	O
component	O
that	O
performs	O
the	O

	O
actions	O
on	O
the	O
victim	O
’s	O
machine	O
.	O
The	O
network	O
component	O
is	O
handled	O
by	O
netid	O
<	O
number>.dll0	O
,	O

	O
which	O
is	O
responsible	O
for	O
receiving	O
commands	O
from	O
the	O
C&C	O
server	O
and	O
sending	O
back	O
their	O

	O
results	O
.	O
When	O
this	O
component	O
receives	O
a	O
command	O
,	O
the	O
command	O
is	O
sent	O
through	O
a	O


	O
14/27	O


	O
localhost	O
socket	O
to	O
winipfile	O
<	O
number>.dll0	O
,	O
which	O
handles	O
the	O
worker	O
component	O
,	O
as	O
shown	O

	O
in	O
Figure	O
3	O
.	O
If	O
initial	O
loopback	O
addresses	O
or	O
ports	O
are	O
in	O
use	O
,	O
both	O
components	O
try	O
to	O
find	O
other	O

	O
available	O
combinations	O
.	O


	O
Figure	O
3	O
.	O
Overall	O
RomCom	O
3.0	O
architecture	O

	O
Bot	O
Commands	O

	O
RomCom	O
3.0	O
commands	O
are	O
received	O
as	O
responses	O
to	O
HTTP	O
POST	O
requests	O
made	O
by	O
the	O

	O
malware	O
network	O
component	O
.	O


	O
Figure	O
4	O
.	O
RomCom	O
3.0	O
command	O
structure	O

	O
15/27	O


	O
Figure	O
4	O
shows	O
an	O
example	O
of	O
command	O
5	O
–	O
a	O
command	O
to	O
download	O
a	O
file	O
to	O
the	O
victim	O
's	O

	O
machine	O
–	O
being	O
received	O
.	O
The	O
ID	O
used	O
for	O
communication	O
is	O
0x950	O
,	O
and	O
command	O
0x05	O
is	O

	O
received	O
with	O
additional	O
data	O
.	O
In	O
this	O
case	O
,	O
the	O
additional	O
data	O
tells	O
the	O
malware	O
running	O
on	O

	O
the	O
infected	O
machine	O
that	O
the	O
downloaded	O
file	O
should	O
occupy	O
939	O
(	O
0x3ac	O
–	O
1	O
)	O
4	O
KB	O
blocks	O
.	O

	O
The	O
file	O
itself	O
is	O
downloaded	O
in	O
a	O
separate	O
response	O
,	O
so	O
in	O
this	O
instance	O
,	O
the	O
final	O
file	O
size	O
on	O

	O
the	O
victim	O
’s	O
side	O
will	O
be	O
3,846,144	O
bytes	O
.	O
As	O
an	O
evasion	O
technique	O
,	O
null	O
bytes	O
are	O
appended	O

	O
to	O
the	O
file	O
to	O
achieve	O
this	O
result	O
.	O
The	O
contents	O
of	O
the	O
additional	O
data	O
field	O
may	O
vary	O
according	O

	O
to	O
the	O
command	O
.	O

	O
In	O
RomCom	O
3.0	O
,	O
we	O
could	O
enumerate	O
42	O
valid	O
commands	O
,	O
as	O
shown	O
in	O
Table	O
2	O
.	O
This	O
is	O
a	O

	O
high	O
number	O
of	O
commands	O
for	O
a	O
regular	O
backdoor	O
,	O
but	O
a	O
few	O
commands	O
are	O
simply	O
variations	O

	O
of	O
others	O
.	O

	O
Command	O


	O
Purpose	O
(	O
from	O
the	O
victim	O
’s	O
perspective	O
)	O


	O
1	O


	O
Send	O
information	O
about	O
connected	O
drives	O


	O
2	O


	O
Send	O
a	O
list	O
of	O
file	O
names	O
under	O
a	O
specified	O
directory	O


	O
3	O


	O
Start	O
cmd.exe	O
to	O
run	O
an	O
existing	O
program	O


	O
4	O


	O
Upload	O
a	O
specified	O
file	O
to	O
the	O
C&C	O
server	O


	O
5	O


	O
Download	O
a	O
file	O
to	O
the	O
victim	O
's	O
machine	O


	O
6	O


	O
Delete	O
a	O
specified	O
file	O
in	O
the	O
victim	O
's	O
machine	O


	O
7	O


	O
Delete	O
a	O
specified	O
directory	O
in	O
the	O
victim	O
's	O
machine	O


	O
8	O


	O
Spawn	O
a	O
given	O
process	O
with	O
PID	O
spoofing	O
(	O
the	O
PID	O
is	O
also	O
given	O
as	O
part	O
of	O

	O
the	O
command	O
data	O
)	O


	O
12	O


	O
Call	O
startWorker	O
(	O
)	O
from	O
%	O
PUBLIC%\Libraries\PhotoDirector.dll	O
,	O
then	O
send	O

	O
%	O
PUBLIC%\Libraries\PhotoDirector.zip	O
to	O
the	O
C&C	O
server	O
and	O
delete	O
it	O


	O
13	O


	O
Call	O
startWorker	O
(	O
)	O
from	O
%	O
PUBLIC%\Libraries\PhotoDirector.dll	O
and	O
write	O

	O
screen	O
information	O
to	O
%	O
PUBLIC%\Libraries\update.conf	O


	O
14	O


	O
Upload	O
%	O
PUBLIC%\Libraries\PhotoDirector.zip	O
to	O
the	O
C&C	O
server	O
and	O
delete	O

	O
it	O


	O
15	O


	O
Send	O
a	O
list	O
of	O
running	O
process	O
with	O
its	O
PIDs	O


	O
16	O


	O
Send	O
a	O
list	O
of	O
installed	O
software	O


	O
17	O


	O
Delete	O
the	O
worker	O
component	O
(	O
winipfile	O
<	O
number>.dll0	O
)	O


	O
18	O


	O
Download	O
a	O
file	O
and	O
save	O
it	O
to	O
%	O
PUBLIC%\Libraries\PhotoDirector.dll	O


	O
16/27	O


	O
19	O


	O
Download	O
a	O
file	O
,	O
save	O
it	O
to	O
%	O
PUBLIC%\Libraries\BrowserData\procsys.dll	O
,	O

	O
and	O
call	O
its	O
stub	O
(	O
)	O
exported	O
function	O


	O
20	O


	O
Download	O
a	O
ZIP	O
archive	O
likely	O
containing	O
3proxy	O
and	O
plink	O
(	O
see	O
command	O
21	O
)	O


	O
21	O


	O
Use	O
3proxy	O
and	O
plink	O
to	O
set	O
up	O
a	O
proxy	O
via	O
SSH	O
.	O
The	O
IP	O
address	O
,	O
password	O
,	O

	O
local	O
,	O
and	O
remote	O
ports	O
are	O
received	O
as	O
command	O
parameters	O
.	O
SSH	O
server	O

	O
username	O
is	O
fixed	O
as	O
“	O
john	O
.	O
”	O


	O
22	O


	O
Kill	O
the	O
3proxy.exe	O
and	O
plink.exe	O
processes	O


	O
23	O


	O
Download	O
a	O
file	O
and	O
save	O
it	O
to	O
%	O
PUBLIC%\Libraries\upd	O
-	O
fil	O
<	O
number>.dll0	O
to	O

	O
update	O
the	O
worker	O


	O
24	O


	O
Send	O
the	O
contents	O
of	O
%	O
PUBLIC%\Libraries\BrowserData\Result	O


	O
25	O


	O
Duplicate	O
the	O
worker	O


	O
26	O


	O
Send	O
the	O
Windows	O
version	O


	O
29	O


	O
Download	O
freeSSHd	O
from	O
the	O
C&C	O
server	O


	O
30	O


	O
Run	O
freeSSHd	O
and	O
use	O
plink	O
to	O
create	O
a	O
reverse	O
connection	O
with	O

	O
51.195.49.215	O
using	O
“	O
john	O
”	O
as	O
the	O
username	O
and	O

	O
“	O
eK6czNHWCT569L1xK9ZH	O
”	O
as	O
the	O
password	O


	O
31	O


	O
Kill	O
the	O
freeSSHd	O
process	O


	O
32	O


	O
Send	O
.txt	O
,	O
.rtf	O
,	O
.xls	O
,	O
.xlsx	O
,	O
.ods	O
,	O
.cmd	O
,	O
.pdf	O
,	O
.vbs	O
,	O
.ps1	O
,	O
.one	O
,	O
.kdb	O
,	O
.kdb	O
,	O
.doc	O
,	O

	O
.doc	O
,	O
.odt	O
,	O
.eml	O
,	O
.msg	O
,	O
and	O
.email	O
files	O
in	O
Downloads	O
,	O
Documents	O
,	O
and	O

	O
Desktop	O
folders	O
under	O
%	O
USERPROFILE%	O


	O
34	O


	O
Run	O
AnyDesk	O
on	O
the	O
victim	O
’s	O
machine	O
on	O
a	O
hidden	O
window	O
and	O
send	O
the	O

	O
AnyDesk	O
ID	O
to	O
the	O
C&C	O
server	O


	O
35	O


	O
Kill	O
the	O
AnyDesk	O
process	O
and	O
delete	O
its	O
executable	O


	O
36	O


	O
Download	O
the	O
AnyDesk	O
executable	O
and	O
save	O
it	O
to	O

	O
%	O
PUBLIC%\Libraries\dsk.exe	O


	O
38	O


	O
Download	O
a	O
file	O
and	O
save	O
it	O
to	O
%	O
PUBLIC%\Libraries\wallet.exe	O


	O
39	O


	O
Download	O
a	O
file	O
and	O
save	O
it	O
to	O
%	O
PUBLIC%\Libraries\7z.dll	O


	O
40	O


	O
Download	O
a	O
file	O
and	O
save	O
it	O
to	O
%	O
PUBLIC%\Libraries\7z.exe	O


	O
41	O


	O
Send	O
the	O
contents	O
of	O
%	O
PUBLIC%\Libraries\tempFolder	O
compressed	O
with	O
7Zip	O


	O
42	O


	O
Download	O
a	O
file	O
and	O
save	O
it	O
to	O
%	O
PUBLIC%\Libraries\7za.exe	O


	O
17/27	O


	O
43	O


	O
Use	O
%	O
PUBLIC%\Libraries\7za.exe	O
to	O
compress	O
a	O
given	O
folder	O
to	O
a	O
fold.zip	O

	O
archive	O
and	O
send	O
the	O
compressed	O
archive	O
to	O
the	O
C&C	O
server	O


	O
44	O


	O
Kill	O
the	O
PhotoDirector.dll	O
process	O


	O
45	O


	O
Download	O
a	O
file	O
and	O
save	O
it	O
to	O
%	O
PUBLIC%\Libraries\msg.dll	O


	O
46	O


	O
Call	O
stW	O
(	O
)	O
function	O
exported	O
by	O
%	O
PUBLIC%\Libraries\msg.dll	O


	O
47	O


	O
Download	O
a	O
file	O
and	O
save	O
it	O
to	O
%	O
PUBLIC%\Libraries\FileInfo.dll	O


	O
48	O


	O
Call	O
fSt	O
(	O
)	O
function	O
exported	O
by	O
%	O
PUBLIC%\Libraries\FileInfo.dll	O


	O
49	O


	O
Update	O
the	O
network	O
component	O


	O
Table	O
2	O
.	O
RomCom	O
3.0	O
commands	O

	O
Additional	O
Malware	O

	O
Based	O
on	O
messages	O
sent	O
back	O
to	O
the	O
C&C	O
server	O
and	O
how	O
the	O
commands	O
use	O
these	O
files	O
,	O

	O
we	O
can	O
infer	O
the	O
purpose	O
of	O
a	O
few	O
additional	O
components	O
:	O

	O
PhotoDirector.dll	O
–	O
a	O
program	O
that	O
takes	O
one	O
or	O
more	O
screenshots	O
and	O
compresses	O

	O
them	O
in	O
a	O
%	O
PUBLIC%\Libraries\PhotoDirector.zip	O
archive	O

	O
procsys.dll	O
–	O
a	O
stealer	O
known	O
as	O
STEALDEAL	O
to	O
retrieve	O
browser	O
cookies	O
and	O
write	O

	O
them	O
to	O
%	O
PUBLIC%\Libraries\BrowserData\Result	O

	O
wallet.exe	O
–	O
a	O
crypto	O
wallet	O
grabber	O
that	O
writes	O
stolen	O
information	O
to	O

	O
%	O
PUBLIC%\Libraries\tempFolder	O

	O
msg.dll	O
–	O
an	O
Instant	O
Messaging	O
grabber	O
to	O
steal	O
chat	O
messages	O

	O
FileInfo.dll	O
–	O
a	O
stealer	O
of	O
FTP	O
credentials	O
,	O
or	O
a	O
component	O
to	O
make	O
the	O
victim	O
’s	O

	O
machine	O
upload	O
files	O
to	O
an	O
FTP	O
server	O

	O
Despite	O
these	O
additional	O
pieces	O
of	O
malware	O
,	O
RomCom	O
3.0	O
also	O
seems	O
to	O
have	O
commands	O
to	O

	O
download	O
and	O
run	O
legitimate	O
software	O
:	O

	O
dsk.exe	O
–	O
a	O
portable	O
version	O
of	O
AnyDesk	O
software	O

	O
7z.dll	O
,	O
7z.exe	O
,	O
and	O
7za.exe	O
–	O
files	O
related	O
to	O
the	O
7	O
-	O
Zip	O
program	O

	O
STEALDEAL	O

	O
The	O
stealer	O
that	O
is	O
downloaded	O
through	O
RomCom	O
’s	O
C&C	O
servers	O
is	O
a	O
relatively	O
simple	O
one	O

	O
that	O
steals	O
stored	O
credentials	O
and	O
browsing	O
history	O
from	O
the	O
following	O
browsers	O
:	O

	O
Google	O
Chrome	O

	O
Microsoft	O
Edge	O

	O
Mozilla	O
Firefox	O


	O
18/27	O


	O
Chromium	O

	O
Chrome	O
Beta	O

	O
Yandex	O
Browser	O

	O
The	O
stealer	O
also	O
collects	O
information	O
on	O
installed	O
mail	O
clients	O
.	O
The	O
stolen	O
data	O
is	O
stored	O
locally	O

	O
on	O
the	O
victim	O
’s	O
machine	O
at	O
%	O
PUBLIC%\Libraries\BrowserData\Result	O
,	O
and	O
through	O
C&C	O

	O
command	O
24	O
,	O
this	O
data	O
is	O
exfiltrated	O
through	O
a	O
RomCom	O
C&C	O
server	O
.	O
We	O
detected	O
the	O

	O
stealer	O
as	O
TrojanSpy	O
.	O
Win64.STEALDEAL	O
,	O
which	O
is	O
also	O
known	O
as	O
SneakyStealer	O
.	O

	O
Evasion	O
Techniques	O

	O
RomCom	O
3.0	O
binaries	O
are	O
protected	O
with	O
VMProtect	O
.	O
Some	O
binaries	O
are	O
also	O
signed	O
with	O

	O
valid	O
certificates	O
.	O
Because	O
the	O
actors	O
decided	O
to	O
use	O
VMProtect	O
’s	O
anti	O
-	O
VM	O
feature	O
,	O
any	O

	O
attempt	O
to	O
run	O
it	O
in	O
a	O
virtual	O
machine	O
(	O
VM	O
)	O
without	O
modification	O
or	O
VM	O
hardening	O
will	O
cause	O

	O
the	O
malware	O
to	O
show	O
an	O
error	O
message	O
and	O
exit	O
(	O
Figure	O
5	O
)	O
.	O


	O
Figure	O
5	O
.	O
Default	O
VMProtect	O
anti	O
-	O
VM	O
detection	O
in	O
RomCom	O
3.0	O
samples	O

	O
Another	O
interesting	O
technique	O
RomCom	O
uses	O
is	O
the	O
ability	O
to	O
add	O
null	O
bytes	O
appended	O
to	O
the	O

	O
files	O
received	O
from	O
a	O
C&C	O
server	O
.	O
Making	O
the	O
file	O
bigger	O
can	O
be	O
an	O
attempt	O
to	O
avoid	O
sandbox	O

	O
products	O
or	O
security	O
software	O
scanners	O
that	O
impose	O
a	O
file	O
size	O
limit	O
.	O

	O
In	O
later	O
versions	O
of	O
RomCom	O
,	O
the	O
binary	O
that	O
is	O
hosted	O
on	O
a	O
lure	O
site	O
contains	O
an	O
encrypted	O

	O
payload	O
.	O
To	O
correctly	O
decrypt	O
the	O
payload	O
,	O
it	O
will	O
need	O
to	O
reach	O
out	O
to	O
a	O
web	O
server	O
at	O
the	O
IP	O

	O
address	O
94.142.138.244	O
and	O
download	O
the	O
decryption	O
key	O
.	O
We	O
suspect	O
this	O
website	O
is	O
a	O

	O
third	O
-	O
party	O
service	O
that	O
is	O
also	O
being	O
used	O
by	O
other	O
malware	O
,	O
including	O
the	O
Vidar	O
stealer	O
that	O
is	O

	O
also	O
known	O
as	O
StealC.	O
Also	O
,	O
recent	O
RomCom	O
droppers	O
have	O
stopped	O
dropping	O
the	O
worker	O

	O
component	O
.	O
Instead	O
,	O
the	O
network	O
component	O
downloads	O
it	O
from	O
the	O
C&C	O
server	O
.	O

	O
Packet	O
Structure	O
and	O
Communications	O
Flow	O


	O
19/27	O


	O
Based	O
on	O
our	O
observations	O
of	O
the	O
communication	O
between	O
victim	O
machines	O
and	O
RomCom	O

	O
C&C	O
servers	O
,	O
we	O
were	O
able	O
to	O
determine	O
what	O
the	O
packet	O
structure	O
of	O
this	O
communication	O

	O
looks	O
like	O
(	O
Figure	O
6	O
)	O
.	O
Initially	O
,	O
the	O
client	O
will	O
reach	O
out	O
to	O
the	O
server	O
with	O
information	O
on	O
the	O

	O
victim	O
’s	O
computer	O
,	O
such	O
as	O
its	O
Universally	O
Unique	O
Identifier	O
(	O
UUID	O
)	O
,	O
username	O
,	O
and	O
the	O

	O
computer	O
name	O
.	O
The	O
server	O
will	O
then	O
respond	O
with	O
a	O
session	O
ID	O
that	O
is	O
four	O
bytes	O
long	O
,	O
as	O

	O
mentioned	O
previously	O
.	O
This	O
session	O
ID	O
is	O
then	O
incremented	O
by	O
one	O
on	O
the	O
first	O
byte	O
by	O
the	O

	O
C&C	O
server	O
with	O
each	O
command	O
that	O
is	O
sent	O
to	O
the	O
victim	O
machine	O
.	O


	O
Figure	O
6	O
.	O
Packet	O
structure	O
of	O
the	O
observed	O
packets	O

	O
One	O
of	O
the	O
first	O
commands	O
we	O
observed	O
was	O
command	O
3	O
,	O
which	O
uses	O
cmd.exe	O
to	O
run	O
a	O

	O
Nltest	O
command	O
with	O
the	O
argument	O
/domain_trusts	O
.	O
This	O
is	O
done	O
to	O
gather	O
information	O
on	O
any	O

	O
domains	O
that	O
the	O
victim	O
machine	O
may	O
know	O
about	O
.	O
Once	O
the	O
command	O
is	O
finished	O
,	O
it	O
returns	O

	O
the	O
results	O
of	O
the	O
command	O
with	O
the	O
Session	O
ID	O
five	O
bytes	O
in	O
;	O
the	O
first	O
four	O
bytes	O
are	O
unknown	O

	O
at	O
this	O
time	O
,	O
but	O
we	O
observed	O
the	O
first	O
byte	O
will	O
be	O
0x01	O
if	O
it	O
is	O
returning	O
data	O
to	O
the	O
server	O
,	O
or	O

	O
0x00	O
if	O
it	O
is	O
receiving	O
data	O
from	O
the	O
server	O
.	O
The	O
C&C	O
server	O
then	O
appears	O
to	O
ask	O
for	O
specific	O

	O
information	O
in	O
an	O
automated	O
manner	O
,	O
as	O
the	O
same	O
requests	O
are	O
sent	O
in	O
quick	O
succession	O

	O
(	O
Figure	O
7	O
)	O
.	O
From	O
our	O
analysis	O
,	O
we	O
have	O
determined	O
that	O
the	O
server	O
is	O
asking	O
for	O
the	O
victim	O

	O
machine	O
to	O
:	O

	O
20/27	O


	O
1	O
.	O
Return	O
ntlest	O
/domain_trusts	O
with	O
command	O
3	O

	O
2	O
.	O
Download	O
StealDeal	O
to	O
collect	O
certain	O
information	O

	O
3	O
.	O
Use	O
StealDeal	O
to	O
collect	O
cookies	O
and	O
other	O
information	O
from	O
the	O
victim	O
’s	O
machine	O

	O
4	O
.	O
Collect	O
files	O
from	O
the	O
Desktop	O
,	O
Documents	O
,	O
and	O
Downloads	O
folders	O
using	O
command	O
32	O


	O
Figure	O
7	O
.	O
Flow	O
of	O
the	O
communication	O
between	O
the	O
C&C	O
server	O
and	O
victim	O
machine	O

	O
Use	O
of	O
fake	O
companies	O
and	O
websites	O

	O
The	O
malware	O
uses	O
certificates	O
to	O
lend	O
credibility	O
to	O
the	O
software	O
that	O
the	O
targeted	O
victims	O

	O
download	O
.	O
On	O
the	O
surface	O
,	O
the	O
companies	O
that	O
are	O
signing	O
these	O
binaries	O
look	O
like	O
legitimate	O

	O
companies	O
that	O
have	O
undergone	O
the	O
process	O
of	O
becoming	O
a	O
signer	O
of	O
these	O
certificates	O
.	O

	O
However	O
,	O
a	O
closer	O
look	O
at	O
these	O
companies	O
’	O
websites	O
reveals	O
several	O
oddities	O
,	O
including	O
nonexistent	O
phone	O
numbers	O
,	O
stock	O
photos	O
of	O
executives	O
,	O
office	O
addresses	O
that	O
do	O
not	O
seem	O
to	O

	O
match	O
.	O
This	O
leads	O
us	O
to	O
believe	O
these	O
are	O
either	O
fake	O
companies	O
or	O
legitimate	O
companies	O
that	O

	O
are	O
being	O
abused	O
in	O
order	O
to	O
pass	O
the	O
checks	O
needed	O
to	O
become	O
an	O
authorized	O
signer	O
of	O

	O
binaries	O
.	O

	O
The	O
RomCom	O
3.0	O
sample	O
that	O
was	O
used	O
in	O
the	O
AstraChat	O
campaign	O
was	O
signed	O
by	O
a	O

	O
Canadian	O
company	O
called	O
Noray	O
Consulting	O
Ltd.	O
,	O
which	O
has	O
a	O
LinkedIn	O
page	O
(	O
Figure	O
8)	O
,	O
a	O

	O
website	O
,	O
and	O
even	O
a	O
listing	O
in	O
a	O
business	O
registry	O
in	O
Canada	O
(	O
Figure	O
9	O
)	O
.	O


	O
21/27	O


	O
Figure	O
8	O
.	O
Screenshot	O
of	O
Noray	O
Consulting	O
's	O
LinkedIn	O
page	O


	O
Figure	O
9	O
.	O
Ontario	O
business	O
registry	O
search	O
results	O
for	O
Noray	O
Consulting	O

	O
The	O
company	O
’s	O
LinkedIn	O
page	O
goes	O
on	O
to	O
mention	O
that	O
Noray	O
Consulting	O
works	O
on	O
SOX	O

	O
compliance	O
,	O
an	O
annual	O
audit	O
mandated	O
by	O
the	O
Sarbanes	O
-	O
Oxley	O
Act	O
(	O
SOX	O
)	O
,	O
as	O
well	O
as	O
other	O

	O
areas	O
of	O
risk	O
control	O
.	O
However	O
,	O
the	O
LinkedIn	O
page	O
also	O
points	O
to	O
a	O
website	O
,	O
noray[.]ca	O
,	O
that	O

	O
does	O
not	O
exist	O
.	O


	O
22/27	O


	O
As	O
the	O
company	O
claims	O
to	O
be	O
based	O
in	O
Ontario	O
according	O
to	O
its	O
LinkedIn	O
page	O
,	O
we	O
looked	O
for	O

	O
any	O
information	O
about	O
it	O
in	O
public	O
records	O
for	O
businesses	O
in	O
Canada	O
.	O
It	O
appears	O
that	O
in	O
2020	O
,	O

	O
the	O
owners	O
of	O
Noray	O
Consulting	O
.	O
Changed	O
the	O
name	O
of	O
the	O
company	O
to	O
just	O
“	O
Noray	O
.	O
”	O
This	O
new	O

	O
company	O
name	O
is	O
not	O
related	O
to	O
any	O
of	O
the	O
things	O
mentioned	O
in	O
this	O
blog	O
post	O
or	O
,	O
from	O
what	O

	O
we	O
can	O
tell	O
,	O
is	O
doing	O
anything	O
malicious	O
.	O
It	O
appears	O
that	O
the	O
actors	O
are	O
watching	O
out	O
for	O

	O
companies	O
that	O
become	O
inactive	O
,	O
or	O
in	O
a	O
similar	O
status	O
,	O
then	O
will	O
appropriate	O
these	O

	O
companies	O
’	O
names	O
.	O

	O
Internet	O
searches	O
for	O
Noray	O
Consulting	O
show	O
that	O
its	O
main	O
website	O
has	O
a	O
non	O
-	O
matching	O

	O
domain	O
name	O
,	O
firstbyteconsulting[.]com	O
.	O
The	O
website	O
used	O
to	O
be	O
for	O
a	O
company	O
that	O

	O
specialized	O
in	O
project	O
management	O
.	O
This	O
domain	O
appears	O
to	O
have	O
expired	O
in	O
2020	O
,	O
but	O
was	O

	O
bought	O
and	O
repurposed	O
to	O
resemble	O
the	O
website	O
from	O
before	O
2020	O
.	O
What	O
ties	O
this	O
domain	O
to	O

	O
Noray	O
Consulting	O
now	O
is	O
that	O
the	O
address	O
details	O
on	O
the	O
website	O
match	O
that	O
which	O
is	O
found	O
on	O

	O
Noray	O
Consulting	O
’s	O
LinkedIn	O
page	O
:	O
a	O
Canadian	O
company	O
in	O
Milton	O
,	O
Ontario	O
.	O
The	O
contact	O
page	O

	O
has	O
a	O
map	O
that	O
shows	O
the	O
company	O
’s	O
location	O
,	O
but	O
the	O
map	O
is	O
in	O
Russian	O
(	O
Figure	O
10	O
)	O
.	O
This	O

	O
could	O
mean	O
that	O
the	O
person	O
who	O
made	O
this	O
Google	O
map	O
had	O
their	O
primary	O
language	O
set	O
to	O

	O
Russian	O
,	O
which	O
would	O
be	O
unusual	O
for	O
a	O
seemingly	O
Canadian	O
-	O
based	O
company	O
.	O


	O
Figure	O
10	O
.	O
Screenshot	O
of	O
the	O
website	O
’s	O
contact	O
page	O
map	O

	O
We	O
also	O
found	O
that	O
the	O
people	O
mentioned	O
on	O
their	O
website	O
are	O
likely	O
stock	O
images	O
or	O
AIgenerated	O
photos	O
of	O
people	O
who	O
are	O
not	O
related	O
in	O
any	O
way	O
to	O
the	O
business	O
,	O
as	O
shown	O
in	O

	O
Figure	O
11	O
.	O


	O
23/27	O


	O
Figure	O
11	O
.	O
Screenshot	O
of	O
the	O
website	O
’s	O
team	O
members	O
page	O

	O
Further	O
investigation	O
reveals	O
that	O
Figure	O
11	O
has	O
a	O
number	O
of	O
red	O
flags	O
:	O

	O
None	O
of	O
these	O
people	O
appear	O
to	O
have	O
real	O
personas	O
on	O
the	O
internet	O

	O
Reverse	O
image	O
search	O
reveals	O
these	O
are	O
stock	O
photo	O
images	O
used	O
on	O
several	O
sites	O

	O
Two	O
members	O
of	O
the	O
team	O
have	O
the	O
same	O
job	O
title	O
of	O
"	O
Manager	O
,	O
HR	O
Process	O
and	O

	O
Compensation	O
"	O

	O
We	O
have	O
also	O
observed	O
that	O
the	O
text	O
in	O
other	O
parts	O
of	O
Noray	O
Consulting	O
’s	O
website	O
has	O
been	O
at	O

	O
least	O
partially	O
copied	O
from	O
other	O
websites	O
.	O
This	O
illustrates	O
that	O
these	O
actors	O
are	O
trying	O
to	O
make	O

	O
the	O
sites	O
believable	O
,	O
offering	O
what	O
seems	O
like	O
realistic	O
services	O
that	O
were	O
lifted	O
from	O
real	O

	O
companies	O
found	O
online	O
.	O

	O
Void	O
Rabisu	O
has	O
had	O
many	O
lure	O
websites	O
that	O
attempt	O
to	O
convince	O
targets	O
to	O
download	O

	O
trojanized	O
legitimate	O
applications	O
.	O
These	O
lure	O
sites	O
look	O
legitimate	O
at	O
first	O
,	O
but	O
usually	O
have	O

	O
similar	O
oddities	O
on	O
the	O
websites	O
.	O
For	O
example	O
,	O
a	O
site	O
that	O
had	O
a	O
business	O
address	O
of	O
a	O

	O
shopping	O
mall	O
,	O
and	O
the	O
contact	O
phone	O
number	O
of	O
a	O
grocery	O
store	O
.	O


	O
24/27	O


	O
Figure	O
12	O
.	O
Screenshot	O
of	O
the	O
contact	O
information	O
of	O
a	O
Void	O
Rabisu	O
lure	O
site	O

	O
We	O
can	O
link	O
the	O
two	O
Canadian	O
companies	O
that	O
were	O
used	O
to	O
sign	O
RomCom	O
binaries	O
in	O
the	O

	O
AstraChat	O
campaign	O
and	O
a	O
campaign	O
against	O
the	O
Ukrainian	O
armed	O
forces	O
with	O
more	O
than	O
80	O

	O
other	O
mostly	O
Canadian	O
companies	O
in	O
total	O
,	O
based	O
on	O
an	O
analysis	O
of	O
internet	O
infrastructure	O
.	O

	O
Among	O
these	O
80	O
other	O
mostly	O
Canadian	O
companies	O
,	O
about	O
two	O
dozen	O
companies	O
were	O
used	O

	O
to	O
sign	O
other	O
malware	O
binaries	O
like	O
Emotet	O
,	O
Matanbuchus	O
Loader	O
,	O
BatLoader	O
,	O
another	O

	O
backdoor	O
known	O
as	O
SolarMarker	O
,	O
and	O
coinminers	O
.	O
This	O
makes	O
us	O
believe	O
that	O
Void	O
Rabisu	O
is	O

	O
likely	O
to	O
be	O
using	O
a	O
third	O
-	O
party	O
service	O
that	O
aids	O
in	O
signing	O
binaries	O
with	O
certificates	O
of	O

	O
seemingly	O
legitimate	O
Canadian	O
companies	O
.	O

	O
Conclusions	O
and	O
Recommendations	O

	O
The	O
war	O
against	O
Ukraine	O
has	O
made	O
cyber	O
campaigns	O
against	O
Ukraine	O
,	O
Eastern	O
Europe	O
,	O
and	O

	O
NATO	O
countries	O
more	O
visible	O
for	O
two	O
reasons	O
:	O
the	O
number	O
of	O
attacks	O
has	O
increased	O

	O
dramatically	O
,	O
and	O
both	O
the	O
private	O
and	O
public	O
sectors	O
are	O
looking	O
closely	O
at	O
what	O
happens	O
in	O

	O
Ukraine	O
.	O
More	O
information	O
from	O
intelligence	O
agencies	O
is	O
being	O
declassified	O
by	O
Western	O

	O
governments	O
,	O
so	O
privately	O
-	O
owned	O
companies	O
can	O
investigate	O
further	O
for	O
themselves	O
.	O
Another	O

	O
important	O
factor	O
is	O
that	O
many	O
actors	O
who	O
previously	O
had	O
different	O
motivations	O
are	O
becoming	O

	O
more	O
aligned	O
towards	O
the	O
same	O
goal	O
,	O
even	O
when	O
their	O
campaigns	O
do	O
not	O
appear	O
to	O
be	O
part	O
of	O

	O
a	O
coordinated	O
effort	O
.	O

	O
The	O
line	O
is	O
blurring	O
between	O
cybercrime	O
driven	O
by	O
financial	O
gain	O
and	O
APT	O
attacks	O
motivated	O

	O
by	O
geopolitics	O
,	O
espionage	O
,	O
disruption	O
,	O
and	O
warfare	O
.	O
Since	O
the	O
rise	O
of	O
Ransomware	O
-	O
as	O
-	O
aService	O
(	O
RaaS	O
)	O
,	O
cybercriminals	O
are	O
now	O
using	O
advanced	O
tactics	O
and	O
targeted	O
attacks	O
that	O

	O
were	O
previously	O
thought	O
to	O
be	O
the	O
domain	O
of	O
APT	O
actors	O
.	O
Inversely	O
,	O
tactics	O
and	O
techniques	O

	O
that	O
were	O
previously	O
used	O
by	O
financially	O
motivated	O
actors	O
are	O
increasingly	O
being	O
used	O
in	O

	O
attacks	O
with	O
geopolitical	O
goals	O
.	O


	O
25/27	O


	O
Currently	O
,	O
APT	O
actors	O
like	O
Pawn	O
Storm	O
and	O
APT29	O
,	O
cyber	O
mercenaries	O
like	O
Void	O
Balaur	O
,	O

	O
hacktivism	O
groups	O
like	O
Killnet	O
,	O
along	O
with	O
cybercriminals	O
like	O
former	O
Conti	O
affiliates	O
and	O
Void	O

	O
Rabisu	O
,	O
are	O
targeting	O
Ukraine	O
and	O
its	O
allies	O
,	O
but	O
their	O
campaigns	O
do	O
not	O
yet	O
look	O
coordinated	O
.	O

	O
We	O
expect	O
that	O
significant	O
geopolitical	O
events	O
like	O
the	O
current	O
war	O
against	O
Ukraine	O
will	O

	O
accelerate	O
the	O
alignment	O
of	O
the	O
campaigns	O
of	O
threat	O
actors	O
who	O
reside	O
in	O
the	O
same	O

	O
geographic	O
region	O
.	O
This	O
will	O
lead	O
to	O
new	O
challenges	O
for	O
defenders	O
,	O
as	O
attacks	O
can	O
then	O
come	O

	O
from	O
many	O
different	O
angles	O
,	O
and	O
it	O
will	O
be	O
less	O
clear	O
who	O
is	O
the	O
actor	O
responsible	O
for	O
them	O
.	O

	O
Based	O
on	O
our	O
analysis	O
,	O
we	O
believe	O
the	O
following	O
activity	O
should	O
be	O
monitored	O
in	O
endpoints	O
:	O

	O
Downloading	O
and	O
executing	O
MSI	O
packages	O
that	O
contain	O
entries	O
in	O
CustomAction	O
tables	O

	O
referring	O
to	O
DLL	O
exported	O
functions	O

	O
Writing	O
access	O
to	O
SOFTWARE\Classes\CLSID\<CSLID	O
>	O
under	O
both	O

	O
HKEY_CURRENT_USER	O
(	O
HKCU	O
)	O
and	O
HKEY_LOCAL_MACHINE	O
(	O
HKLM	O
)	O
,	O
which	O
can	O

	O
be	O
a	O
sign	O
of	O
COM	O
hijacking	O

	O
Initiation	O
of	O
localhost	O
sockets	O
by	O
rundll32.exe	O
,	O
as	O
RomCom	O
DLLs	O
are	O
loaded	O
by	O
this	O

	O
process	O
—	O
we	O
observed	O
that	O
RomCom	O
listens	O
on	O
the	O
port	O
range	O
5554	O
-	O
5600	O
when	O

	O
setting	O
up	O
localhost	O
sockets	O

	O
Binary	O
padding	O
with	O
null	O
bytes	O
,	O
a	O
known	O
technique	O
to	O
evade	O
scanners	O
.	O
Although	O

	O
RomCom	O
did	O
n’t	O
use	O
this	O
feature	O
in	O
our	O
tests	O
,	O
it	O
is	O
present	O
in	O
command	O
5	O
.	O
We	O
included	O
a	O

	O
YARA	O
ruleset	O
to	O
look	O
for	O
such	O
files	O
in	O
our	O
GitHub	O
research	O
repository	O
.	O

	O
Binary	O
padding	O
with	O
non	O
-	O
zero	O
data	O
,	O
which	O
we	O
observed	O
in	O
one	O
sample	O
when	O
it	O
was	O

	O
dropping	O
another	O
.	O
This	O
alone	O
is	O
not	O
malicious	O
,	O
but	O
it	O
is	O
worth	O
flagging	O
once	O
detected	O
for	O

	O
further	O
investigation	O
.	O

	O
Endpoint	O
solutions	O
like	O
Trend	O
Micro	O
's	O
Smart	O
Protection	O
Suites	O
and	O
Worry	O
-	O
Free	O
Business	O

	O
Security	O
solutions	O
also	O
offer	O
protection	O
for	O
both	O
users	O
and	O
businesses	O
against	O
threats	O
like	O

	O
RomCom	O
.	O
These	O
solutions	O
come	O
equipped	O
with	O
behavior	O
-	O
monitoring	O
capabilities	O
that	O
enable	O

	O
them	O
to	O
detect	O
malicious	O
files	O
,	O
scripts	O
,	O
and	O
messages	O
.	O
They	O
can	O
also	O
block	O
all	O
related	O

	O
malicious	O
URLs	O
.	O
Additionally	O
,	O
the	O
Trend	O
Micro	O
™	O
Deep	O
Discovery	O
™	O
solution	O
includes	O
an	O
email	O

	O
inspection	O
layer	O
that	O
can	O
identify	O
and	O
protect	O
enterprises	O
from	O
malicious	O
attachments	O
and	O

	O
URLs	O
.	O
By	O
leveraging	O
these	O
powerful	O
tools	O
,	O
users	O
and	O
businesses	O
can	O
effectively	O
defend	O

	O
themselves	O
against	O
the	O
damaging	O
effects	O
of	O
RomCom	O
and	O
other	O
similar	O
threats	O
.	O

	O
Indicators	O
of	O
Compromise	O

	O
Download	O
the	O
full	O
list	O
of	O
indicators	O
here	O
.	O

	O
MITRE	O
ATT&CK	O

	O
ID	O


	O
Name	O


	O
Description	O


	O
26/27	O


	O
T1583.008	O


	O
Acquire	O
Infrastructure	O
:	O

	O
Malvertising	O


	O
RomCom	O
uses	O
malvertising	O
to	O
redirect	O
targets	O

	O
to	O
lure	O
websites	O
from	O
which	O
to	O
download	O
fake	O

	O
installer	O
applications	O


	O
T1566.002	O


	O
Phishing	O
:	O
Spear	O
Phishing	O

	O
Link	O


	O
RomCom	O
sent	O
highly	O
targeted	O
spear	O
phishing	O

	O
emails	O


	O
T1027.002	O


	O
Obfuscated	O
Files	O
or	O

	O
Information	O
:	O
Software	O

	O
Packing	O


	O
RomCom	O
uses	O
VMProtect	O


	O
T1027.001	O


	O
Obfuscated	O
Files	O
or	O

	O
Information	O
:	O
Binary	O

	O
Padding	O


	O
RomCom	O
uses	O
binary	O
padding	O
on	O
dropped	O
files	O

	O
to	O
avoid	O
security	O
solutions	O


	O
T1546.015	O


	O
Event	O
Triggered	O
Execution	O
:	O

	O
Component	O
Object	O
Model	O

	O
Hijacking	O


	O
RomCom	O
uses	O
COM	O
hijacking	O
for	O
persistence	O


	O
T1571	O


	O
Non	O
-	O
Standard	O
Port	O


	O
RomCom	O
listens	O
on	O
port	O
ranges	O
5554	O
to	O
5600	O

	O
for	O
communication	O
between	O
dropped	O

	O
components	O


	O
T1071.001	O


	O
Application	O
Layer	O
Protocol	O
:	O

	O
Web	O
Protocols	O


	O
RomCom	O
uses	O
HTTPS	O
for	O
C&C	O

	O
communications	O


	O
T1555.003	O


	O
Credentials	O
from	O
Password	O

	O
Stores	O
:	O
Credentials	O
from	O

	O
Web	O
Browsers	O


	O
RomCom	O
uses	O
a	O
stealer	O
to	O
gather	O
credentials	O

	O
of	O
several	O
browsers	O


	O
T1113	O


	O
Screen	O
Capture	O


	O
RomCom	O
can	O
capture	O
screenshots	O
of	O
the	O

	O
victim	O
's	O
machine	O


	O
T1219	O


	O
Remote	O
Access	O
Software	O


	O
RomCom	O
's	O
backdoor	O
has	O
a	O
functionality	O
to	O
run	O

	O
AnyDesk	O
application	O


	O
27/27	O