conll/text_31.conll

Sophisticated	O
APT29	O
Campaign	O
Abuses	O
Notion	O
API	O
to	O

	O
Target	O
the	O
European	O
Commission	O

	O
mrtiepolo.medium.com/sophisticated-apt29-campaign-abuses-notion-api-to-target-the-european-commission200188059f58	O

	O
Gianluca	O
Tiepolo	O


	O
March	O
30	O
,	O
2023	O


	O
Gianluca	O
Tiepolo	O


	O
1/9	O


	O
Mar	O
9	O

	O
·	O

	O
12	O
min	O
read	O

	O
Research	O
by	O
Gianluca	O
Tiepolo	O

	O
A	O
map	O
of	O
Russia	O
,	O
as	O
imagined	O
by	O
DALL·E	O


	O
APT29	O
is	O
a	O
highly	O
sophisticated	O
Advanced	O
Persistent	O
Threat	O
(	O
APT	O
)	O
group	O
that	O
has	O
been	O

	O
attributed	O
to	O
Russia	O
’s	O
Foreign	O
Intelligence	O
Service	O
(	O
SVR	O
)	O
.	O
The	O
group	O
has	O
been	O
active	O

	O
since	O
at	O
least	O
2008	O
and	O
has	O
been	O
involved	O
in	O
a	O
wide	O
range	O
of	O
espionage	O
and	O
cyber	O
-	O
attack	O

	O
campaigns	O
targeting	O
governments	O
,	O
military	O
organizations	O
,	O
defense	O
contractors	O
,	O
and	O

	O
various	O
industries	O
in	O
the	O
United	O
States	O
,	O
Europe	O
,	O
and	O
Asia	O
.	O

	O
APT29	O
is	O
also	O
known	O
as	O
NOBELIUM	O
(	O
Microsoft	O
)	O
,	O
Cozy	O
Bear	O
(	O
Crowdstrike	O
)	O
,	O
The	O
Dukes	O

	O
(	O
Kaspersky	O
)	O
,	O
JACKMACKEREL	O
(	O
iDefense	O
)	O
,	O
BlueBravo	O
(	O
Recorded	O
Future	O
)	O
and	O
UNC2452	O

	O
(	O
FireEye	O
)	O
.	O

	O
The	O
group	O
is	O
known	O
for	O
its	O
subtle	O
and	O
sophisticated	O
tradecraft	O
in	O
stealing	O
geopolitical	O

	O
intelligence	O
:	O
unlike	O
other	O
Russian	O
state	O
-	O
sponsored	O
groups	O
such	O
as	O
APT28	O
or	O
Sandworm	O
,	O

	O
APT29	O
has	O
not	O
been	O
linked	O
to	O
destructive	O
operations	O
and	O
operates	O
with	O
much	O
more	O

	O
discretion	O
.	O

	O
The	O
group	O
has	O
been	O
attributed	O
to	O
a	O
number	O
of	O
high	O
-	O
profile	O
cyberattacks	O
,	O
including	O
:	O

	O
APT29	O
was	O
one	O
of	O
the	O
two	O
Russian	O
groups	O
responsible	O
for	O
the	O
cyberattack	O
on	O
the	O
DNC	O

	O
during	O
the	O
2016	O
U.S.	O
presidential	O
election	O
.	O
The	O
group	O
gained	O
access	O
to	O
the	O
DNC	O
’s	O

	O
email	O
system	O
and	O
stole	O
sensitive	O
information	O
,	O
which	O
was	O
subsequently	O
leaked	O
to	O
the	O

	O
public	O
.	O

	O
APT29	O
was	O
attributed	O
to	O
the	O
highly	O
sophisticated	O
supply	O
chain	O
attack	O
against	O

	O
SolarWinds	O
,	O
a	O
leading	O
IT	O
management	O
software	O
provider	O
.	O
The	O
attack	O
allowed	O
the	O
group	O

	O
to	O
gain	O
access	O
to	O
the	O
systems	O
of	O
several	O
U.S.	O
government	O
agencies	O
,	O
including	O
the	O

	O
Department	O
of	O
Justice	O
,	O
the	O
Department	O
of	O
State	O
,	O
and	O
the	O
Department	O
of	O
Homeland	O

	O
Security	O
.	O

	O
In	O
May	O
2021	O
,	O
it	O
was	O
revealed	O
that	O
APT29	O
was	O
responsible	O
for	O
a	O
large	O
-	O
scale	O
cyberattack	O
on	O

	O
multiple	O
U.S.	O
government	O
agencies	O
and	O
private	O
companies	O
,	O
including	O
Microsoft	O
.	O
The	O
group	O

	O
used	O
a	O
compromised	O
email	O
marketing	O
system	O
to	O
send	O
spear	O
-	O
phishing	O
emails	O
to	O
over	O
3,000	O

	O
individual	O
accounts	O
,	O
resulting	O
in	O
the	O
installation	O
of	O
a	O
backdoor	O
that	O
allowed	O
the	O
attackers	O
to	O


	O
2/9	O


	O
gain	O
access	O
to	O
the	O
victims	O
’	O
networks	O
.	O
The	O
group	O
has	O
also	O
been	O
linked	O
to	O
other	O
significant	O

	O
cyberattacks	O
,	O
including	O
the	O
theft	O
of	O
COVID-19	O
research	O
from	O
U.S.-based	O
pharmaceutical	O

	O
companies	O
.	O

	O
Overall	O
,	O
APT29	O
is	O
one	O
of	O
the	O
most	O
sophisticated	O
and	O
well	O
-	O
resourced	O
APT	O
groups	O
in	O
the	O
world	O
,	O

	O
and	O
its	O
TTPs	O
are	O
constantly	O
evolving	O
and	O
changing	O
.	O


	O
Tactics	O
,	O
Techniques	O
,	O
and	O
Procedures	O

	O
APT29	O
is	O
known	O
for	O
its	O
patient	O
and	O
persistent	O
targeting	O
of	O
its	O
victims	O
,	O
often	O
using	O
multi	O
-	O
stage	O

	O
attacks	O
that	O
take	O
weeks	O
or	O
even	O
months	O
to	O
complete	O
.	O
Following	O
is	O
a	O
list	O
of	O
the	O
group	O
’s	O
most	O

	O
notable	O
TTPs	O
:	O

	O
the	O
group	O
uses	O
highly	O
targeted	O
and	O
convincing	O
spear	O
-	O
phishing	O
emails	O
to	O
gain	O
access	O
to	O

	O
a	O
target	O
’s	O
system	O
.	O
These	O
emails	O
are	O
usually	O
tailored	O
to	O
the	O
recipient	O
’s	O
interests	O
and	O

	O
appear	O
to	O
come	O
from	O
a	O
trusted	O
source	O
.	O

	O
APT29	O
has	O
also	O
been	O
known	O
to	O
use	O
watering	O
hole	O
attacks	O
,	O
where	O
the	O
group	O

	O
compromises	O
a	O
trusted	O
website	O
frequented	O
by	O
the	O
target	O
,	O
and	O
then	O
injects	O
malware	O
into	O

	O
the	O
site	O
to	O
infect	O
visitors	O
.	O

	O
APT29	O
uses	O
highly	O
customized	O
malware	O
,	O
such	O
as	O
“	O
SeaDuke	O
”	O
and	O
“	O
CosmicDuke	O
,	O
”	O
that	O

	O
are	O
designed	O
to	O
evade	O
detection	O
and	O
maintain	O
persistence	O
on	O
the	O
target	O
system	O
.	O
The	O

	O
group	O
is	O
also	O
known	O
to	O
use	O
well	O
-	O
known	O
tools	O
like	O
“	O
Cobalt	O
Strike	O
”	O
and	O
“	O
PowerShell	O

	O
Empire	O
”	O
.	O

	O
the	O
group	O
is	O
known	O
for	O
its	O
use	O
of	O
zero	O
-	O
day	O
exploits	O
to	O
gain	O
access	O
to	O
target	O
systems	O
.	O
For	O

	O
example	O
,	O
APT29	O
has	O
been	O
known	O
to	O
use	O
exploits	O
for	O
popular	O
software	O
like	O
Microsoft	O

	O
Office	O
and	O
Adobe	O
Flash	O
.	O

	O
APT29	O
often	O
uses	O
“	O
”	O
tactics	O
,	O
where	O
the	O
group	O
uses	O
legitimate	O
tools	O
and	O
techniques	O
that	O

	O
are	O
already	O
present	O
on	O
the	O
target	O
system	O
to	O
evade	O
detection	O
.	O
This	O
can	O
include	O
tools	O
like	O

	O
,	O
,	O
and	O
.	O

	O
In	O
this	O
particular	O
research	O
,	O
I	O
focused	O
on	O
analyzing	O
APT29	O
’s	O
command	O
-	O
and	O
-	O
control	O

	O
capabilities	O
.	O


	O
Command	O
&	O
Control	O

	O
This	O
threat	O
group	O
has	O
a	O
history	O
of	O
using	O
trusted	O
and	O
legitimate	O
cloud	O
services	O
(	O
such	O
as	O

	O
social	O
media	O
services	O
and	O
Google	O
Drive	O
)	O
for	O
their	O
cyber	O
attacks	O
in	O
an	O
attempt	O
to	O
blend	O
into	O

	O
normal	O
network	O
traffic	O
and	O
evade	O
detection	O
.	O
Malware	O
distributed	O
by	O
APT29	O
also	O
contains	O

	O
the	O
ability	O
to	O
exfiltrate	O
data	O
over	O
those	O
same	O
C2	O
channels	O
.	O
For	O
example	O
:	O

	O
The	O
group	O
’s	O
malware	O
searched	O
for	O
specific	O
that	O
contained	O
URLs	O
to	O
access	O
C2	O
servers	O
.	O

	O
APT29	O
’s	O
and	O
malware	O
also	O
have	O
the	O
ability	O
to	O
use	O
to	O
obtain	O
C2	O
URLs	O
,	O
as	O
well	O
as	O
other	O

	O
social	O
media	O
services	O
like	O
and	O


	O
3/9	O


	O
APT29	O
’s	O
backdoor	O
uses	O
,	O
,	O
and	O
for	O
C2	O
communication	O
.	O

	O
APT29	O
has	O
also	O
utilized	O
custom	O
encryption	O
methods	O
,	O
such	O
as	O
those	O
found	O
in	O
the	O
group	O
’s	O

	O
SeaDuke	O
malware	O
where	O
a	O
unique	O
fingerprint	O
was	O
generated	O
for	O
the	O
infected	O
host	O
and	O

	O
Base64	O
encoding	O
and	O
RC4	O
/	O
AES	O
encryption	O
was	O
used	O
to	O
layer	O
data	O
during	O
communications	O

	O
with	O
their	O
C2	O
server	O
.	O
The	O
group	O
has	O
also	O
employed	O
techniques	O
such	O
as	O
“	O
domain	O
fronting	O
”	O
and	O

	O
TOR	O
obfuscation	O
plugins	O
to	O
create	O
encrypted	O
network	O
tunnels	O
.	O

	O
Using	O
social	O
networks	O
for	O
C2	O
communications	O
is	O
not	O
an	O
entirely	O
new	O
technique	O
:	O
other	O

	O
Russian	O
groups	O
such	O
as	O
Turla	O
(	O
Venomous	O
Bear	O
)	O
leveraged	O
comments	O
posted	O
to	O
Instagram	O

	O
to	O
obtain	O
the	O
address	O
of	O
its	O
command	O
and	O
control	O
servers	O
.	O

	O
Source	O
:	O
FireEye	O
,	O
Stealthy	O
Tactics	O
Define	O
a	O
Russian	O
Cyber	O
Threat	O
Group	O
,	O
2015	O

	O
APT29	O
was	O
spotted	O
using	O
Twitter	O
to	O
control	O
infected	O
machines	O
as	O
early	O
as	O
2015	O
:	O
in	O
the	O

	O
HAMMERTOSS	O
campaign	O
,	O
the	O
group	O
was	O
able	O
to	O
receive	O
commands	O
and	O
send	O
stolen	O
data	O

	O
through	O
the	O
popular	O
social	O
network	O
,	O
which	O
allowed	O
them	O
to	O
evade	O
detection	O
by	O
security	O

	O
solutions	O
that	O
did	O
not	O
monitor	O
social	O
media	O
traffic	O
.	O


	O
EnvyScout	O

	O
In	O
a	O
more	O
recent	O
campaign	O
dating	O
back	O
to	O
June	O
2021	O
,	O
APT29	O
targeted	O
Italy	O
diplomatic	O

	O
organizations	O
with	O
a	O
spear	O
phishing	O
campaign	O
that	O
distributed	O
the	O
EnvyScout	O
backdoor	O
.	O

	O
C2	O
communication	O
through	O
Slack	O

	O
The	O
backdoor	O
first	O
calls	O
a	O
function	O
to	O
create	O
a	O
custom	O
Slack	O
channel	O
,	O
adding	O
the	O
attacker	O
’s	O

	O
user	O
ID	O
to	O
the	O
newly	O
created	O
channel	O
.	O
The	O
backdoor	O
gets	O
the	O
user	O
name	O
and	O
hostname	O
of	O
the	O

	O
victim	O
host	O
,	O
adds	O
4	O
random	O
numbers	O
to	O
form	O
the	O
name	O
of	O
the	O
channel	O
,	O
and	O
sends	O
an	O
HTTP	O

	O
request	O
with	O
an	O
authorization	O
token	O
to	O
the	O
Slack	O
API	O
.	O
After	O
the	O
channel	O
is	O
established	O
,	O
the	O

	O
backdoor	O
enters	O
an	O
infinite	O
loop	O
:	O
it	O
uses	O
the	O
“	O
chat.postMessage	O
”	O
API	O
request	O
to	O
send	O
a	O

	O
beacon	O
message	O
to	O
the	O
newly	O
created	O
channel	O
and	O
it	O
receives	O
a	O
response	O
with	O
a	O
list	O
of	O

	O
additional	O
files	O
and	O
payloads	O
that	O
are	O
downloaded	O
and	O
executed	O
on	O
the	O
target	O
machine	O
.	O


	O
Beatdrop	O

	O
In	O
mid	O
-	O
January	O
2022	O
,	O
APT29	O
launched	O
yet	O
another	O
spear	O
phishing	O
campaign	O
targeting	O
a	O

	O
diplomatic	O
entity	O
,	O
which	O
was	O
detected	O
and	O
responded	O
to	O
by	O
Mandiant	O
.	O
During	O
the	O

	O
investigation	O
,	O
Mandiant	O
discovered	O
that	O
the	O
malicious	O
emails	O
were	O
used	O
to	O
distribute	O
the	O

	O
BEATDROP	O
and	O
BOOMMIC	O
downloaders	O
.	O

	O
BEATDROP	O
is	O
a	O
downloader	O
written	O
in	O
C	O
that	O
leverages	O
Trello	O
for	O
Command	O
-	O
and	O
-	O
Control	O

	O
(	O
C2	O
)	O
communication	O
.	O
Trello	O
is	O
a	O
web	O
-	O
based	O
project	O
management	O
application	O
that	O
allows	O

	O
users	O
to	O
organize	O
tasks	O
and	O
projects	O
using	O
customizable	O
boards	O
,	O
lists	O
,	O
and	O
cards	O
.	O


	O
4/9	O


	O
When	O
executed	O
,	O
BEATDROP	O
maps	O
its	O
own	O
copy	O
of	O
ntdll.dll	O
into	O
memory	O
to	O
execute	O

	O
shellcode	O
in	O
its	O
own	O
process	O
.	O
It	O
creates	O
a	O
suspended	O
thread	O
,	O
then	O
enumerates	O
the	O
system	O
for	O

	O
the	O
username	O
,	O
computer	O
name	O
,	O
and	O
IP	O
address	O
to	O
create	O
a	O
victim	O
ID	O
.	O
This	O
victim	O
ID	O
is	O
used	O

	O
by	O
BEATDROP	O
to	O
store	O
and	O
retrieve	O
victim	O
payloads	O
from	O
its	O
C2	O
.	O
Once	O
the	O
victim	O
ID	O
is	O

	O
created	O
,	O
BEATDROP	O
sends	O
an	O
initial	O
request	O
to	O
Trello	O
to	O
determine	O
if	O
the	O
current	O
victim	O
has	O

	O
already	O
been	O
compromised	O
.	O
The	O
shellcode	O
payload	O
is	O
then	O
retrieved	O
from	O
Trello	O
and	O
is	O

	O
targeted	O
for	O
each	O
victim	O
.	O
Once	O
the	O
payload	O
has	O
been	O
retrieved	O
,	O
it	O
is	O
deleted	O
from	O
Trello	O
.	O


	O
Notion	O
for	O
C2	O
Communication	O

	O
In	O
October	O
2022	O
,	O
ESET	O
Research	O
discovered	O
a	O
sample	O
uploaded	O
to	O
VirusTotal	O
that	O
closely	O

	O
resembled	O
what	O
APT29	O
had	O
used	O
a	O
few	O
months	O
ago	O
,	O
with	O
the	O
key	O
difference	O
being	O
that	O
it	O

	O
used	O
Notion	O
,	O
a	O
cloud	O
-	O
based	O
note	O
-	O
taking	O
software	O
platform	O
,	O
for	O
Command	O
-	O
and	O
-	O
Control	O

	O
(	O
C&C	O
)	O
communications	O
.	O

	O
Notion	O
API	O
can	O
be	O
abused	O
for	O
C2	O
communications	O
by	O
embedding	O
the	O
commands	O
into	O
the	O

	O
Notion	O
workspace	O
,	O
which	O
is	O
accessed	O
by	O
the	O
malware	O
as	O
if	O
it	O
were	O
a	O
legitimate	O
user	O
.	O
This	O

	O
misuse	O
of	O
Notion	O
allows	O
the	O
threat	O
actors	O
to	O
evade	O
detection	O
and	O
bypass	O
security	O

	O
controls	O
,	O
as	O
the	O
traffic	O
between	O
the	O
malware	O
and	O
the	O
Notion	O
server	O
is	O
likely	O
to	O
be	O
perceived	O

	O
as	O
legitimate	O
traffic	O
.	O

	O
ESET	O
researchers	O
suspect	O
that	O
the	O
downloader	O
deployed	O
in	O
this	O
particular	O
campaign	O
was	O

	O
designed	O
to	O
gather	O
and	O
execute	O
additional	O
malicious	O
payloads	O
,	O
such	O
as	O
Cobalt	O
Strike	O
.	O
The	O

	O
campaign	O
has	O
been	O
analyzed	O
in	O
more	O
detail	O
by	O
researchers	O
at	O
Hive	O
Pro	O
and	O
Recorded	O

	O
Future	O
,	O
which	O
identify	O
the	O
sample	O
as	O
the	O
GraphicalNeutrino	O
malware	O
.	O

	O
According	O
to	O
Recorded	O
Future	O
,	O
APT29	O
utilized	O
a	O
compromised	O
website	O
with	O
a	O
lure	O
text	O
of	O

	O
“	O
Ambassador	O
’s	O
schedule	O
November	O
2022	O
”	O
to	O
distribute	O
the	O
ZIP	O
file	O
“	O
schedule.zip	O
”	O
,	O
suggesting	O

	O
that	O
the	O
targets	O
of	O
the	O
campaign	O
are	O
related	O
to	O
embassy	O
staff	O
or	O
an	O
ambassador	O
.	O

	O
GraphicalNeutrino	O
,	O
the	O
malware	O
used	O
in	O
the	O
operation	O
,	O
serves	O
as	O
a	O
loader	O
with	O
basic	O
C2	O

	O
capabilities	O
and	O
employs	O
various	O
anti	O
-	O
analysis	O
techniques	O
to	O
avoid	O
detection	O
,	O
including	O
API	O

	O
unhooking	O
and	O
sandbox	O
evasion	O
.	O

	O
GraphicalNeutrino	O
artifact	O
—	O
140runtime.dll	O

	O
After	O
establishing	O
persistence	O
,	O
the	O
malware	O
decrypts	O
several	O
strings	O
,	O
including	O
a	O
Notion	O
API	O

	O
key	O
and	O
a	O
database	O
identifier	O
,	O
and	O
calculates	O
a	O
unique	O
ID	O
for	O
the	O
victim	O
based	O
on	O
their	O

	O
username	O
and	O
hostname	O
.	O
It	O
then	O
uses	O
Notion	O
’s	O
API	O
for	O
C2	O
communication	O
to	O
deliver	O

	O
additional	O
payloads	O
to	O
the	O
victim	O
’s	O
machine	O
.	O

	O
For	O
each	O
request	O
to	O
the	O
C2	O
,	O
GraphicalNeutrino	O
parses	O
the	O
JSON	O
-	O
formatted	O
response	O
and	O

	O
searches	O
for	O
a	O
“	O
file	O
”	O
array	O
;	O
if	O
the	O
array	O
is	O
not	O
empty	O
,	O
then	O
the	O
malware	O
will	O
parse	O
out	O
the	O
URL	O

	O
field	O
,	O
download	O
the	O
file	O
and	O
decrypt	O
it	O
using	O
a	O
custom	O
cipher	O
.	O
Once	O
the	O
shellcode	O
is	O

	O
decrypted	O
,	O
it	O
is	O
indirectly	O
spawned	O
in	O
a	O
new	O
thread	O
.	O


	O
5/9	O


	O
A	O
sample	O
response	O
from	O
Notion	O
C2	O

	O
The	O
use	O
of	O
diplomatic	O
lures	O
during	O
times	O
of	O
heightened	O
geopolitical	O
tensions	O
,	O
such	O
as	O
the	O

	O
ongoing	O
war	O
in	O
Ukraine	O
,	O
is	O
likely	O
to	O
be	O
effective	O
for	O
Russian	O
APT	O
groups	O
,	O
given	O
the	O
potential	O

	O
impact	O
of	O
information	O
gathered	O
from	O
compromised	O
entities	O
or	O
individuals	O
on	O
Russian	O

	O
foreign	O
policy	O
and	O
strategic	O
decision	O
-	O
making	O
processes	O
.	O
It	O
is	O
perhaps	O
for	O
this	O
reason	O
that	O

	O
APT29	O
adopted	O
the	O
same	O
tactics	O
—	O
in	O
particular	O
the	O
stealthy	O
C2	O
communication	O
through	O

	O
Notion	O
—	O
for	O
its	O
next	O
big	O
campaign	O
,	O
this	O
time	O
targeting	O
the	O
European	O
Commission	O
.	O


	O
Attack	O
against	O
the	O
European	O
Commission	O

	O
In	O
this	O
final	O
section	O
of	O
the	O
blog	O
post	O
,	O
I	O
’m	O
dissecting	O
a	O
previously	O
undisclosed	O
campaign	O

	O
attributed	O
to	O
APT29	O
which	O
targeted	O
the	O
European	O
Commission	O
.	O
The	O
previous	O
introduction	O
to	O

	O
the	O
group	O
’s	O
TTPs	O
and	O
campaigns	O
will	O
hopefully	O
be	O
beneficial	O
to	O
the	O
reader	O
,	O
as	O
this	O
attack	O

	O
shares	O
quite	O
a	O
few	O
similarities	O
with	O
the	O
GraphicalNeutrino	O
campaign	O
that	O
was	O
exposed	O
by	O

	O
Recorded	O
Future	O
.	O


	O
Initial	O
Access	O

	O
Beginning	O
mid	O
-	O
February	O
2023	O
,	O
a	O
spear	O
phishing	O
campaign	O
targeted	O
a	O
number	O
of	O
email	O

	O
addresses	O
related	O
to	O
members	O
of	O
the	O
European	O
Commission	O
.	O
The	O
attack	O
involved	O
the	O

	O
distribution	O
of	O
a	O
malicious	O
.iso	O
image	O
that	O
contained	O
a	O
new	O
sample	O
of	O
the	O
VaporRage	O

	O
downloader	O
.	O
Once	O
executed	O
,	O
the	O
malware	O
was	O
observed	O
exploiting	O
the	O
Notion	O
API	O
to	O
deploy	O

	O
Cobalt	O
Strike	O
beacons	O
.	O

	O
Execution	O
flow	O
for	O
the	O
attack	O
targeting	O
the	O
European	O
Commission	O

	O
The	O
first	O
phishing	O
email	O
,	O
sent	O
on	O
the	O
13th	O
February	O
2023	O
,	O
masqueraded	O
as	O
an	O
administrative	O

	O
notice	O
related	O
to	O
documents	O
available	O
to	O
download	O
from	O
eTrustEx	O
,	O
a	O
web	O
based	O
exchange	O

	O
platform	O
that	O
ensures	O
secure	O
transmission	O
of	O
documents	O
between	O
members	O
of	O
the	O

	O
Commission	O
.	O
The	O
decoy	O
emails	O
are	O
written	O
in	O
English	O
and	O
were	O
delivered	O
to	O
an	O
extremely	O

	O
targeted	O
number	O
of	O
key	O
people	O
that	O
use	O
the	O
eTrustEx	O
platform	O
.	O

	O
Lure	O
email	O
delivered	O
to	O
the	O
European	O
Commission	O

	O
In	O
addition	O
,	O
I	O
noticed	O
that	O
in	O
different	O
samples	O
of	O
the	O
email	O
,	O
the	O
senders	O
are	O
probably	O

	O
compromised	O
email	O
accounts	O
belonging	O
to	O
legitimate	O
government	O
organizations	O
.	O
This	O
could	O

	O
lead	O
victims	O
to	O
believe	O
that	O
the	O
emails	O
came	O
from	O
reliable	O
partners	O
,	O
making	O
it	O
more	O
likely	O
for	O

	O
recipients	O
to	O
click	O
on	O
the	O
links	O
.	O

	O
When	O
the	O
link	O
is	O
opened	O
,	O
the	O
victim	O
is	O
redirected	O
to	O
a	O
malicious	O
HTML	O
page	O
hosted	O
at	O

	O
hxxps://literaturaelsalvador[.]com	O
/	O
Instructions.html	O
which	O
makes	O
use	O
of	O
a	O

	O
technique	O
known	O
as	O
HTML	O
Smuggling	O
to	O
download	O
an	O
ISO	O
image	O
to	O
the	O
target	O
system	O
.	O
I	O

	O
believe	O
that	O
this	O
domain	O
is	O
not	O
actor	O
-	O
owned	O
but	O
has	O
been	O
compromised	O
,	O
which	O
aligns	O
with	O

	O
previous	O
APT29	O
activity	O
.	O


	O
6/9	O


	O
Lure	O
website	O

	O
The	O
ISO	O
file	O
is	O
set	O
to	O
auto	O
-	O
download	O
when	O
the	O
website	O
is	O
visited	O
by	O
the	O
victim	O
;	O
this	O
is	O

	O
achieved	O
through	O
the	O
following	O
JavaScript	O
code	O
.	O
The	O
contents	O
of	O
Instructions.iso	O
is	O
stored	O

	O
in	O
the	O
d	O
variable	O
.	O

	O
JS	O
Code	O
which	O
downloads	O
the	O
first	O
-	O
stage	O
payload	O


	O
Execution	O

	O
Once	O
the	O
file	O
has	O
been	O
written	O
to	O
disk	O
,	O
when	O
a	O
user	O
double	O
-	O
clicks	O
on	O
it	O
in	O
Windows	O
10	O
or	O
later	O
,	O

	O
the	O
image	O
is	O
mounted	O
and	O
the	O
folder	O
contents	O
is	O
displayed	O
in	O
Windows	O
Explorer	O
.	O
The	O
ISO	O

	O
contains	O
two	O
files	O
—	O
a	O
Windows	O
shortcut	O
file	O
(	O
Instructions.lnk	O
)	O
and	O
a	O
malicious	O
DLL	O

	O
(	O
BugSplatRc64.dll	O
)	O
.	O

	O
If	O
the	O
user	O
clicks	O
on	O
the	O
LNK	O
file	O
,	O
the	O
following	O
command	O
runs	O
,	O
unintentionally	O
triggering	O
the	O

	O
execution	O
of	O
the	O
malicious	O
DLL	O
.	O

	O
Execution	O
of	O
the	O
malicious	O
DLL	O

	O
Using	O
LNK	O
shortcuts	O
to	O
execute	O
malicious	O
DLLs	O
is	O
a	O
technique	O
that	O
has	O
been	O
associated	O
to	O

	O
APT29	O
in	O
a	O
number	O
of	O
campaigns	O
.	O
In	O
this	O
particular	O
scenario	O
,	O
I	O
recognized	O
the	O
sample	O
as	O

	O
VaporRage	O
,	O
a	O
downloader	O
that	O
has	O
been	O
used	O
by	O
APT29	O
since	O
2021	O
.	O


	O
Persistence	O

	O
When	O
executed	O
with	O
the	O
InitiateDs	O
export	O
,	O
VaporRage	O
first	O
runs	O
a	O
few	O
reconnaissance	O

	O
commands	O
and	O
generates	O
a	O
host	O
-	O
id	O
by	O
hex	O
-	O
encoding	O
the	O
DNS	O
domain	O
and	O
username	O
.	O
Then	O
,	O

	O
it	O
creates	O
a	O
copy	O
of	O
itself	O
at	O
:	O

	O
C:\Users\%USERNAME%\AppData\Local\DsDiBacks\BugSplatRc64.dll	O


	O
VaporRage	O
creates	O
a	O
copy	O
of	O
itself	O

	O
VaporRage	O
then	O
establishes	O
persistence	O
on	O
the	O
compromised	O
system	O
by	O
creating	O
a	O
registry	O

	O
run	O
key	O
located	O
at	O
:	O
\Software\Microsoft\Windows\CurrentVersion\Run\DsDiBacks	O
.	O

	O
VaporRage	O
establishes	O
persistence	O
through	O
a	O
registry	O
key	O


	O
Command	O
and	O
Control	O

	O
As	O
I	O
anticipated	O
at	O
the	O
start	O
of	O
this	O
post	O
,	O
the	O
VaporRage	O
sample	O
delivered	O
in	O
the	O
execution	O

	O
chain	O
leverages	O
its	O
command	O
-	O
and	O
-	O
control	O
by	O
communicating	O
over	O
HTTPS	O
using	O
Notion	O

	O
APIs	O
.	O
Notion	O
’s	O
database	O
feature	O
is	O
also	O
used	O
to	O
store	O
victim	O
information	O
and	O
stage	O
further	O

	O
payloads	O
for	O
download	O
.	O

	O
PCAP	O
collected	O
during	O
C2	O
communication	O


	O
7/9	O


	O
Based	O
on	O
my	O
observations	O
,	O
this	O
VaporRage	O
sample	O
periodically	O
executes	O
a	O
POST	O
request	O
to	O

	O
the	O
Notion	O
API	O
to	O
check	O
the	O
availability	O
of	O
a	O
second	O
-	O
stage	O
malware	O
payload	O
,	O
which	O
is	O
then	O

	O
retrieved	O
and	O
executed	O
in	O
memory	O
.	O
In	O
this	O
particular	O
campaign	O
,	O
APT29	O
used	O
VaporRage	O
to	O

	O
distribute	O
Cobalt	O
Strike	O
beacons	O
to	O
further	O
establish	O
a	O
foothold	O
within	O
the	O
environment	O
.	O

	O
Following	O
is	O
a	O
sample	O
POST	O
request	O
towards	O
api.notion.com	O
(	O
104[.]18.42.99	O
):	O

	O
POST	O
/v1	O
/	O
databases/37089abc0926463182bb5343bce252cc	O
/	O
query	O
HTTP/1.1	O

	O
content	O
-	O
type	O
:	O
application	O
/	O
json	O

	O
accept	O
:	O
application	O
/	O
json	O

	O
notion	O
-	O
version	O
:	O
2022	O
-	O
06	O
-	O
28	O

	O
authorization	O
:	O
Bearer	O
secret_X92sXCVWoTk63aPgGKlPBBmHVmuKXJ2geugKa7Ogj7s	O

	O
User	O
-	O
Agent	O
:	O
Mozilla/5.0	O
(	O
Windows	O
NT	O
6.3	O
;	O
WOW64	O
)	O
AppleWebKit/537.36	O
(	O
KHTML	O
,	O
like	O

	O
Gecko	O
)	O
Chrome/35.0.1916.114	O
Safari/537.36	O

	O
Host	O
:	O
api.notion.com	O

	O
Content	O
-	O
Length	O
:	O
79	O

	O
Connection	O
:	O
Keep	O
-	O
Alive	O

	O
Cache	O
-	O
Control	O
:	O
no	O
-	O
cache	O

	O
{	O
"	O
filter":{"property":"Name","rich_text":{"equals":"VKoMr3830"}},"page_size":1	O
}	O


	O
This	O
technique	O
exemplifies	O
APT29	O
’s	O
ongoing	O
attempts	O
to	O
obscure	O
its	O
actions	O
and	O
maintain	O

	O
continuous	O
access	O
to	O
target	O
systems	O
.	O
This	O
has	O
been	O
documented	O
thoroughly	O
by	O
Mandiant	O
,	O

	O
who	O
have	O
described	O
APT29	O
using	O
a	O
variety	O
of	O
techniques	O
,	O
including	O
scheduled	O
tasks	O
,	O
run	O

	O
keys	O
,	O
malicious	O
certificates	O
,	O
and	O
in	O
-	O
memory	O
backdoors	O
,	O
sometimes	O
utilizing	O
multiple	O

	O
methods	O
for	O
each	O
target	O
.	O

	O
Overall	O
,	O
the	O
use	O
of	O
cloud	O
services	O
such	O
as	O
Trello	O
and	O
Notion	O
for	O
C2	O
communications	O
not	O

	O
only	O
provides	O
a	O
threat	O
actor	O
with	O
increased	O
capabilities	O
for	O
evasion	O
of	O
network	O
security	O

	O
controls	O
,	O
but	O
also	O
increases	O
resilience	O
to	O
law	O
enforcement	O
takedowns	O
:	O
social	O
media	O
and	O
cloud	O

	O
services	O
are	O
often	O
hosted	O
on	O
multiple	O
servers	O
and	O
locations	O
,	O
making	O
it	O
more	O
difficult	O
for	O

	O
authorities	O
to	O
take	O
down	O
the	O
entire	O
platform	O
.	O
This	O
means	O
that	O
the	O
threat	O
actor	O
can	O
continue	O
to	O

	O
use	O
the	O
platform	O
for	O
C2	O
communications	O
even	O
if	O
some	O
servers	O
are	O
taken	O
down	O
.	O
These	O

	O
advantages	O
make	O
it	O
an	O
attractive	O
option	O
for	O
threat	O
groups	O
such	O
as	O
APT29	O
to	O
conduct	O
their	O

	O
malicious	O
activities	O
.	O


	O
Conclusions	O

	O
The	O
range	O
of	O
tactics	O
,	O
techniques	O
,	O
and	O
procedures	O
(	O
TTPs	O
)	O
used	O
by	O
APT29	O
in	O
this	O
campaign	O

	O
supports	O
the	O
conclusion	O
that	O
their	O
objective	O
is	O
to	O
establish	O
numerous	O
means	O
of	O
long	O
-	O
term	O

	O
access	O
to	O
facilitate	O
espionage	O
-	O
related	O
intelligence	O
gathering	O
within	O
the	O
targeted	O

	O
government	O
entities	O
’	O
victim	O
networks	O
.	O
Nations	O
that	O
have	O
a	O
connection	O
to	O
the	O
Ukraine	O
crisis	O
,	O

	O
specifically	O
those	O
with	O
significant	O
geopolitical	O
,	O
economic	O
,	O
or	O
military	O
ties	O
to	O
Russia	O
or	O
Ukraine	O
,	O

	O
face	O
a	O
heightened	O
risk	O
of	O
being	O
targeted	O
by	O
APT29	O
.	O


	O
8/9	O


	O
This	O
threat	O
group	O
has	O
shown	O
an	O
impressive	O
ability	O
to	O
adapt	O
swiftly	O
during	O
their	O
operations	O
.	O

	O
They	O
use	O
innovative	O
and	O
unique	O
methods	O
to	O
circumvent	O
detection	O
and	O
authentication	O

	O
requirements	O
in	O
their	O
target	O
environments	O
.	O
In	O
their	O
recent	O
operations	O
,	O
the	O
group	O
has	O

	O
demonstrated	O
a	O
deep	O
understanding	O
of	O
operational	O
security	O
,	O
enabling	O
them	O
to	O
move	O

	O
seamlessly	O
between	O
on	O
-	O
premises	O
and	O
cloud	O
resources	O
with	O
minimal	O
use	O
of	O
malware	O
.	O
These	O

	O
factors	O
,	O
combined	O
with	O
their	O
advanced	O
malware	O
development	O
skills	O
,	O
long	O
history	O
of	O

	O
operations	O
,	O
and	O
extended	O
time	O
on	O
targets	O
,	O
indicate	O
that	O
APT29	O
is	O
a	O
well	O
-	O
funded	O
and	O

	O
exceptionally	O
sophisticated	O
actor	O
and	O
will	O
definitely	O
continue	O
to	O
be	O
a	O
threat	O
during	O
2023	O
.	O


	O
IOCs	O

	O
Following	O
is	O
a	O
list	O
of	O
indicators	O
associated	O
to	O
this	O
campaign	O
.	O

	O
Domains	O
hxxps://literaturaelsalvador[.]com	O
/	O
instructions.html	O

	O
hxxps://api[.]notion[.]com	O
/	O
v1	O
/	O
databases/37089abc0926463182bb5343bce252cc	O
/	O
query	O

	O
IPs	O

	O
108[.]167.180.186	O

	O
104[.]18.42.99	O

	O
Files	O
—	O
SHA256	O

	O
21a0b617431850a9ea2698515c277cbd95de4e59c493d0d8f194f3808eb16354	O

	O
(	O
Instructions.iso	O
)	O

	O
e957326b2167fa7ccd508cbf531779a28bfce75eb2635ab81826a522979aeb98	O

	O
(	O
BugSplatRc64.dll	O
)	O

	O
About	O
the	O
Author	O
—	O
Gianluca	O
Tiepolo	O

	O
I	O
’m	O
a	O
cybersecurity	O
researcher	O
who	O
specializes	O
in	O
digital	O
forensics	O
and	O
incident	O
response	O
for	O

	O
the	O
telecommunications	O
industry	O
.	O
Over	O
the	O
past	O
12	O
years	O
,	O
by	O
working	O
as	O
a	O
consultant	O
I	O
have	O

	O
performed	O
forensic	O
analysis	O
,	O
threat	O
hunting	O
,	O
incident	O
response	O
,	O
and	O
Cyber	O
Threat	O
Intelligence	O

	O
analysis	O
for	O
dozens	O
of	O
organizations	O
,	O
including	O
several	O
Fortune	O
®	O
100	O
companies	O
.	O
In	O
2013	O
,	O
I	O

	O
co	O
-	O
founded	O
the	O
startup	O
Sixth	O
Sense	O
Solutions	O
,	O
which	O
developed	O
AI	O
-	O
based	O
antifraud	O
solutions	O
.	O

	O
Today	O
,	O
I	O
work	O
as	O
a	O
Cyber	O
Threat	O
Intelligence	O
(	O
CTI	O
)	O
Team	O
Lead	O
for	O
Accenture	O
Security	O
.	O

	O
I	O
love	O
writing	O
and	O
sharing	O
my	O
knowledge	O
:	O
in	O
2016	O
I	O
authored	O
the	O
book	O
“	O
Getting	O
Started	O
with	O

	O
RethinkDB	O
”	O
,	O
and	O
in	O
2022	O
I	O
wrote	O
“	O
iOS	O
Forensics	O
for	O
Investigators	O
”	O
,	O
both	O
published	O
by	O
Packt	O

	O
Publishing	O
.	O


	O
9/9	O