Earth	O
Preta	O
Spear	O
-	O
Phishing	O
Governments	O
Worldwide	O

	O
trendmicro.com/en_us/research/22/k/earth-preta-spear-phishing-governments-worldwide.html	O

	O
November	O
18	O
,	O
2022	O


	O
APT	O
&	O
Targeted	O
Attacks	O

	O
We	O
break	O
down	O
the	O
cyberespionage	O
activities	O
of	O
advanced	O
persistent	O
threat	O
(	O
APT	O
)	O
group	O

	O
Earth	O
Preta	O
,	O
observed	O
in	O
large	O
-	O
scale	O
attack	O
deployments	O
that	O
began	O
in	O
March	O
.	O
We	O
also	O

	O
show	O
the	O
infection	O
routines	O
of	O
the	O
malware	O
families	O
they	O
use	O
to	O
infect	O
multiple	O
sectors	O

	O
worldwide	O
:	O
TONEINS	O
,	O
TONESHELL	O
,	O
and	O
PUBLOAD	O
.	O

	O
By	O
:	O
Nick	O
Dai	O
,	O
Vickie	O
Su	O
,	O
Sunny	O
Lu	O
November	O
18	O
,	O
2022	O
Read	O
time	O
:	O
(	O
words	O
)	O

	O
We	O
have	O
been	O
monitoring	O
a	O
wave	O
of	O
spear	O
-	O
phishing	O
attacks	O
targeting	O
the	O
government	O
,	O

	O
academic	O
,	O
foundations	O
,	O
and	O
research	O
sectors	O
around	O
the	O
world	O
.	O
Based	O
on	O
the	O
lure	O

	O
documents	O
we	O
observed	O
in	O
the	O
wild	O
,	O
this	O
is	O
a	O
large	O
-	O
scale	O
cyberespionage	O
campaign	O
that	O

	O
began	O
around	O
March	O
.	O
After	O
months	O
of	O
tracking	O
,	O
the	O
seemingly	O
wide	O
outbreak	O
of	O
targeted	O

	O
attacks	O
includes	O
but	O
not	O
limited	O
to	O
Myanmar	O
,	O
Australia	O
,	O
the	O
Philippines	O
,	O
Japan	O
and	O
Taiwan	O
.	O

	O
We	O
analyzed	O
the	O
malware	O
families	O
used	O
in	O
this	O
campaign	O
and	O
attributed	O
the	O
incidents	O
to	O
a	O

	O
notorious	O
advanced	O
persistent	O
threat	O
(	O
APT	O
)	O
group	O
called	O
Earth	O
Preta	O
(	O
also	O
known	O
as	O

	O
Mustang	O
Panda	O
and	O
Bronze	O
President	O
)	O
.	O


	O
1/30	O


	O
Figure	O
1	O
.	O
Country	O
distribution	O
of	O
Earth	O
Preta	O
attacks	O
from	O
May	O
to	O
October	O
2022	O

	O
In	O
our	O
observation	O
of	O
the	O
campaigns	O
,	O
we	O
noted	O
that	O
,	O
Earth	O
Preta	O
abused	O
fake	O
Google	O

	O
accounts	O
to	O
distribute	O
the	O
malware	O
via	O
spear	O
-	O
phishing	O
emails	O
,	O
initially	O
stored	O
in	O
an	O
archive	O
file	O

	O
(	O
such	O
as	O
rar	O
/	O
zip	O
/	O
jar	O
)	O
and	O
distributed	O
through	O
Google	O
Drive	O
links	O
.	O
Users	O
are	O
then	O
lured	O
into	O

	O
downloading	O
and	O
triggering	O
the	O
malware	O
to	O
execute	O
,	O
TONEINS	O
,	O
TONESHELL	O
,	O
and	O

	O
PUBLOAD	O
.	O
PUBLOAD	O
has	O
been	O
previously	O
reported	O
,	O
but	O
we	O
add	O
new	O
technical	O
insights	O
in	O

	O
this	O
entry	O
that	O
tie	O
it	O
to	O
TONEINS	O
and	O
TONESHELL	O
,	O
newly	O
discovered	O
malware	O
families	O
used	O

	O
by	O
the	O
group	O
for	O
its	O
campaigns	O
.	O

	O
In	O
addition	O
,	O
the	O
actors	O
leverage	O
different	O
techniques	O
for	O
evading	O
detection	O
and	O
analysis	O
,	O
like	O

	O
code	O
obfuscation	O
and	O
custom	O
exception	O
handlers	O
.	O
We	O
also	O
found	O
that	O
the	O
senders	O
of	O
the	O

	O
spear	O
-	O
phishing	O
emails	O
and	O
the	O
owners	O
of	O
Google	O
Drive	O
links	O
are	O
the	O
same	O
.	O
Based	O
on	O
the	O

	O
sample	O
documents	O
that	O
were	O
used	O
for	O
luring	O
the	O
victims	O
,	O
we	O
also	O
believe	O
that	O
the	O
attackers	O

	O
were	O
able	O
to	O
conduct	O
research	O
and	O
,	O
potentially	O
,	O
prior	O
breaches	O
on	O
the	O
target	O
organizations	O

	O
that	O
allowed	O
for	O
familiarity	O
,	O
as	O
indicated	O
in	O
the	O
abbreviation	O
of	O
names	O
from	O
previously	O

	O
compromised	O
accounts	O
.	O

	O
In	O
this	O
blog	O
entry	O
,	O
we	O
discuss	O
Earth	O
Preta	O
’s	O
new	O
campaign	O
and	O
its	O
tactics	O
,	O
techniques	O
,	O
and	O

	O
procedures	O
(	O
TTPs	O
)	O
,	O
including	O
new	O
installers	O
and	O
backdoors	O
.	O
Last	O
,	O
we	O
share	O
how	O
security	O

	O
practitioners	O
can	O
track	O
malware	O
threats	O
similar	O
to	O
those	O
that	O
we	O
have	O
identified	O
.	O

	O
Initial	O
compromise	O
and	O
targets	O


	O
2/30	O


	O
Based	O
on	O
our	O
monitoring	O
of	O
this	O
threat	O
,	O
the	O
decoy	O
documents	O
are	O
written	O
in	O
Burmese	O
,	O
and	O

	O
the	O
contents	O
are	O
"	O
လျှို့ဝှက်ချက်	O
"	O
(	O
“	O
Internal	O
-	O
only	O
”	O
)	O
.	O
Most	O
of	O
the	O
topics	O
in	O
the	O

	O
documents	O
are	O
controversial	O
issues	O
between	O
countries	O
and	O
contain	O
words	O
like	O
"	O
Secret	O
"	O
or	O

	O
“	O
Confidential	O
.	O
”	O
These	O
could	O
indicate	O
that	O
the	O
attackers	O
are	O
targeting	O
Myanmar	O
government	O

	O
entities	O
as	O
their	O
first	O
entry	O
point	O
.	O
This	O
could	O
also	O
mean	O
that	O
the	O
attackers	O
have	O
already	O

	O
compromised	O
specific	O
political	O
entities	O
prior	O
to	O
the	O
attack	O
,	O
something	O
that	O
Talos	O
Intelligence	O

	O
had	O
also	O
previously	O
noted	O
.	O

	O
The	O
attackers	O
use	O
the	O
stolen	O
documents	O
as	O
decoys	O
to	O
trick	O
the	O
targeted	O
organizations	O

	O
working	O
with	O
Myanmar	O
government	O
offices	O
into	O
downloading	O
and	O
executing	O
the	O
malicious	O

	O
files	O
.	O
The	O
victimology	O
covers	O
a	O
broad	O
range	O
of	O
organizations	O
and	O
verticals	O
worldwide	O
,	O
with	O
a	O

	O
higher	O
concentration	O
in	O
the	O
Asia	O
Pacific	O
region	O
.	O
Apart	O
from	O
the	O
government	O
offices	O
with	O

	O
collaborative	O
work	O
in	O
Myanmar	O
,	O
subsequent	O
victims	O
included	O
the	O
education	O
and	O
research	O

	O
industries	O
,	O
among	O
others	O
.	O
In	O
addition	O
to	O
decoy	O
topics	O
covering	O
ongoing	O
international	O
events	O

	O
concerning	O
specific	O
organizations	O
,	O
the	O
attackers	O
also	O
lure	O
individuals	O
with	O
subject	O
headings	O

	O
pertaining	O
to	O
pornographic	O
materials	O
.	O


	O
Figure	O
2	O
.	O
Distribution	O
of	O
Earth	O
Preta	O
’s	O
targeted	O
industries	O

	O
Analyzing	O
the	O
routines	O


	O
3/30	O


	O
Figure	O
3	O
.	O
Earth	O
Preta	O
attack	O
campaign	O
routine	O
from	O
March	O
to	O
October	O
2022	O

	O
Earth	O
Preta	O
uses	O
spear	O
-	O
phishing	O
emails	O
as	O
its	O
first	O
step	O
for	O
intrusion	O
.	O
As	O
aforementioned	O
,	O

	O
some	O
of	O
the	O
emails	O
’	O
subjects	O
and	O
contents	O
discuss	O
geopolitical	O
topics	O
,	O
while	O
others	O
might	O

	O
contain	O
sensational	O
subjects	O
.	O
We	O
observed	O
that	O
all	O
the	O
emails	O
we	O
analyzed	O
had	O
the	O
Google	O

	O
Drive	O
links	O
embedded	O
in	O
them	O
,	O
which	O
points	O
to	O
how	O
users	O
might	O
be	O
tricked	O
into	O
downloading	O

	O
the	O
malicious	O
archives	O
.	O
The	O
file	O
types	O
of	O
the	O
archives	O
include	O
compressed	O
files	O
such	O
as	O
.rar	O
,	O

	O
.zip	O
,	O
and	O
.jar	O
,	O
to	O
name	O
a	O
few	O
.	O
Upon	O
accessing	O
the	O
links	O
,	O
we	O
learned	O
that	O
the	O
archives	O
contain	O

	O
the	O
malware	O
TONEINS	O
,	O
TONESHELL	O
,	O
and	O
PUBLOAD	O
malware	O
families	O
.	O


	O
4/30	O


	O
Figure	O
4	O
.	O


	O
Email	O
document	O
sample	O
of	O
meeting	O
minutes	O
,	O
likely	O
stolen	O
from	O
a	O
prior	O
compromise	O

	O
Spear	O
-	O
phishing	O
emails	O

	O
We	O
analyzed	O
the	O
contents	O
of	O
the	O
emails	O
and	O
observed	O
that	O
a	O
Google	O
Drive	O
link	O
is	O
used	O
as	O
a	O

	O
lure	O
for	O
victims	O
.	O
The	O
email	O
's	O
subject	O
might	O
be	O
empty	O
or	O
might	O
have	O
the	O
same	O
name	O
as	O
the	O

	O
malicious	O
archive	O
.	O
Rather	O
than	O
add	O
the	O
victims	O
’	O
addresses	O
to	O
the	O
email	O
’s	O
“	O
To	O
”	O
header	O
,	O
the	O

	O
5/30	O


	O
threat	O
actors	O
used	O
fake	O
emails	O
.	O
Meanwhile	O
,	O
the	O
real	O
victims	O
'	O
addresses	O
were	O
written	O
in	O
the	O

	O
"	O
CC	O
"	O
header	O
,	O
likely	O
to	O
evade	O
security	O
analysis	O
and	O
slow	O
down	O
investigations	O
.	O
Using	O
opensource	O
intelligence	O
(	O
OSINT	O
)	O
tool	O
GHunt	O
to	O
probe	O
those	O
Gmail	O
addresses	O
in	O
the	O
“	O
To	O
”	O
section	O
,	O

	O
we	O
found	O
these	O
fake	O
accounts	O
with	O
little	O
information	O
in	O
them	O
.	O

	O
Moreover	O
,	O
we	O
observed	O
that	O
some	O
of	O
the	O
senders	O
might	O
be	O
compromised	O
email	O
accounts	O

	O
from	O
a	O
specific	O
organization	O
.	O
Victims	O
might	O
be	O
convinced	O
that	O
these	O
mails	O
were	O
sent	O
from	O

	O
trusted	O
partners	O
,	O
increasing	O
the	O
chances	O
that	O
recipients	O
will	O
select	O
the	O
malicious	O
links	O
.	O

	O
Decoy	O
documents	O

	O
We	O
also	O
found	O
some	O
decoy	O
documents	O
linked	O
to	O
the	O
organizations	O
related	O
to	O
or	O
working	O
with	O

	O
Myanmar	O
government	O
entities	O
.	O
The	O
first	O
decoy	O
's	O
file	O
name	O
is	O
Assistance	O
and	O

	O
Recovery(china).exe	O
,	O
while	O
another	O
decoy	O
.PDF	O
document	O
(	O
“	O
ပြည်ထောင်စုသမ္မတ	O

	O
မြန်မာနိုင်ငံတော်သံရုံး.pdf	O
,	O
meaning	O
“	O
Embassy	O
of	O
the	O
Republic	O
of	O
Myanmar	O
"	O
)	O
was	O
observed	O

	O
in	O
a	O
compressed	O
file	O
named	O
Assistance	O
and	O
Recovery(china).rar	O
.	O
Allegedly	O
,	O
this	O
is	O
a	O

	O
document	O
containing	O
the	O
ambassador	O
’s	O
report	O
in	O
rough	O
meeting	O
schedules	O
between	O
the	O

	O
embassies	O
of	O
Myanmar	O
and	O
China	O
.	O

	O
Another	O
document	O
is	O
related	O
to	O
the	O
Japan	O
Society	O
for	O
the	O
Promotion	O
of	O
Science	O
(	O
JSPS	O
)	O
,	O
an	O

	O
initiative	O
that	O
provides	O
researchers	O
opportunities	O
to	O
conduct	O
and	O
undergo	O
research	O

	O
exchanges	O
in	O
Japan	O
.	O
Notably	O
,	O
the	O
documents	O
in	O
the	O
compressed	O
file	O
attachment(EN).rar	O
are	O

	O
mostly	O
image	O
files	O
.	O
The	O
malicious	O
DLL	O
and	O
the	O
executable	O
,	O
which	O
are	O
used	O
for	O
the	O
next	O
layer	O

	O
of	O
sideloading	O
,	O
are	O
also	O
included	O
among	O
them	O
.	O


	O
6/30	O


	O
Figure	O
5	O
.	O
Sample	O
decoy	O
documents	O
relating	O
to	O
government	O
meetings	O
(	O
left	O
)	O
and	O
overseas	O

	O
research	O
exchange	O
(	O
right	O
)	O

	O
There	O
are	O
also	O
other	O
decoy	O
documents	O
with	O
diverse	O
content	O
themes	O
,	O
including	O
regional	O
affairs	O

	O
and	O
pornography	O
.	O
However	O
,	O
when	O
the	O
victim	O
opens	O
the	O
fake	O
document	O
file	O
in	O
this	O
folder	O
,	O
no	O

	O
corresponding	O
content	O
appears	O
.	O

	O
Arrival	O
vectors	O

	O
We	O
observed	O
at	O
least	O
three	O
types	O
of	O
arrival	O
vectors	O
as	O
the	O
intrusions	O
'	O
entry	O
points	O
,	O
including	O

	O
over	O
30	O
lure	O
archives	O
around	O
the	O
world	O
distributed	O
via	O
Google	O
Drive	O
links	O
,	O
Dropbox	O
links	O
,	O
or	O

	O
other	O
IP	O
addresses	O
hosting	O
the	O
files	O
.	O
In	O
most	O
of	O
the	O
archives	O
we	O
collected	O
,	O
there	O
are	O
legitimate	O

	O
executables	O
,	O
as	O
well	O
as	O
the	O
sideloaded	O
DLL	O
.	O
The	O
names	O
of	O
the	O
archives	O
and	O
the	O
decoy	O

	O
documents	O
vary	O
in	O
each	O
case	O
.	O
In	O
the	O
following	O
sections	O
,	O
we	O
take	O
some	O
of	O
them	O
as	O
examples	O

	O
and	O
share	O
the	O
TTPs	O
of	O
each	O
.	O

	O
Type	O
A	O
:	O
DLL	O
sideloading	O

	O
In	O
this	O
case	O
,	O
there	O
are	O
three	O
files	O
in	O
the	O
archive	O
:	O
"	O
~	O
,	O
"	O
Increasingly	O
confident	O
US	O
is	O
baiting	O

	O
China.exe	O
,	O
and	O
libcef.dll	O
.	O
Notably	O
,	O
the	O
names	O
of	O
the	O
lure	O
documents	O
and	O
executables	O
can	O
be	O

	O
different	O
,	O
as	O
detailed	O
in	O
the	O
next	O
sections	O
.	O

	O
7/30	O


	O
Filename	O


	O
Detection	O


	O
220509	O
-	O
(	O
Cabinet	O
Meeting	O

	O
2022).zip	O


	O
~	O


	O
Increasingly	O
confident	O
US	O
is	O

	O
baiting	O
China.exe	O

	O
libcef.dll	O


	O
Description	O

	O
Lure	O

	O
document	O

	O
Legitimate	O

	O
executable	O
for	O
DLL	O

	O
sideloading	O


	O
Trojan	O
.	O
Win32.PUBLOAD	O


	O
Malicious	O
DLL	O


	O
Table	O
1	O
.	O
Files	O
in	O
the	O
archive	O
of	O
Type	O
A	O


	O
8/30	O


	O
Figure	O
6	O
.	O
An	O
example	O
of	O
a	O
decoy	O
document	O
from	O
the	O
PUBLOAD	O
archives	O

	O
Inside	O
the	O
archive	O
,	O
the	O
"	O
~	O
"	O
file	O
is	O
a	O
lure	O
document	O
.	O
The	O
executable	O
Increasingly	O
confident	O
US	O

	O
is	O
baiting	O
China.exe	O
is	O
a	O
legitimate	O
executable	O
(	O
originally	O
named	O

	O
adobe_licensing_wf_helper.exe	O
,	O
which	O
is	O
the	O
Adobe	O
Licensing	O
WF	O
Helper	O
)	O
.	O
This	O
executable	O

	O
will	O
sideload	O
the	O
malicious	O
libcef.dll	O
and	O
trigger	O
the	O
export	O
function	O
cef_api_hash	O
.	O

	O
When	O
executed	O
for	O
the	O
first	O
time	O
,	O
the	O
executable	O
tries	O
to	O
install	O
the	O
malware	O
by	O
copying	O
the	O

	O
.exe	O
file	O
and	O
moving	O
libcef.dll	O
(	O
detected	O
by	O
Trend	O
Micro	O
as	O
Trojan	O
.	O
Win32.PUBLOAD	O
)	O
to	O

	O
<	O
%	O
PUBLIC%	O
>	O
Both	O
.exe	O
and	O
.dll	O
files	O
will	O
be	O
renamed	O

	O
C:\Users\Public\Pictures\adobe_wf.exe	O
and	O
C:\Users\Public\Pictures\libcef.dll	O
,	O
respectively	O
.	O

	O
Additionally	O
,	O
"	O
~	O
"	O
is	O
renamed	O
as	O
05-09-2022.docx	O
and	O
dropped	O
to	O
the	O
Desktop	O
.	O


	O
9/30	O


	O
Figure	O
7	O
.	O
Type	O
A	O
’s	O
malicious	O
routine	O

	O
Type	O
B	O
:	O
Shortcut	O
links	O

	O
The	O
malicious	O
archive	O
contains	O
three	O
files	O
:	O
New	O
Word	O
Document.lnk	O
,	O
putty.exe	O
,	O
and	O

	O
CefBrowser.dll	O
.	O
In	O
particular	O
,	O
the	O
DLL	O
and	O
executable	O
files	O
are	O
placed	O
in	O
multiple	O
layers	O
of	O

	O
folders	O
named	O
“	O
_	O
”	O
.	O

	O
Filename	O


	O
Detection	O


	O
Desktop.rar	O


	O
New	O
Word	O
Document.lnk	O


	O
_	O
\_\_\_\_\_\putty.exe	O


	O
_	O
\_\_\_\_\_\CefBrowser.dll	O


	O
Description	O

	O
Installer	O

	O
Legitimate	O

	O
executable	O
for	O
DLL	O

	O
sideloading	O


	O
Backdoor	O
.	O
Win32.TONESHELL	O


	O
Malicious	O
DLL	O


	O
Table	O
2	O
.	O
Files	O
in	O
the	O
archive	O
of	O
Type	O
B	O

	O
The	O
threat	O
actor	O
utilizes	O
the	O
.lnk	O
file	O
to	O
install	O
the	O
malicious	O
files	O
by	O
decompressing	O
the	O
archive	O

	O
file	O
with	O
WinRAR	O
.	O
The	O
full	O
command	O
line	O
is	O
as	O
follows	O
.	O

	O
%	O
ComSpec%	O
/c	O
"	O
_	O
\_\_\_\_\_\putty.exe||(forfiles	O
/P	O
%	O
APPDATA%\	O
..	O
\	O
..	O
\	O
/S	O
/M	O
Desktop.rar	O

	O
/C	O
"	O
cmd	O
/c	O
(	O
c:\progra~1\winrar\winrar.exe	O
x	O
-inul	O
-o+	O
@path||c:\progra~2\winrar\winrar.exe	O
x	O
inul	O
-o+	O
@path)&&_\_\_\_\_\_\putty.exe	O
"	O
)	O
"	O


	O
10/30	O


	O
Putty.exe	O
is	O
masquerading	O
as	O
a	O
normal	O
executable	O
;	O
its	O
original	O
file	O
name	O
is	O
AppXUpdate.exe	O
.	O

	O
When	O
it	O
is	O
executed	O
,	O
it	O
sideloads	O
CefBrowser.dll	O
and	O
executes	O
the	O
main	O
routine	O
in	O
its	O
export	O

	O
function	O
,	O
CCefInterface::SubProcessMain	O
.	O
It	O
also	O
abuses	O
schtasks	O
for	O
persistence	O
.	O


	O
Figure	O
8	O
.	O
Type	O
B	O
's	O
malicious	O
routine	O

	O
Type	O
C	O
:	O
Fake	O
file	O
extensions	O

	O
In	O
this	O
case	O
,	O
China	O
VS	O
Taiwan.rar	O
contains	O
several	O
files	O
,	O
including	O
:	O

	O
Filename	O


	O
Detection	O


	O
China	O
VS	O
Taiwan.rar	O


	O
China	O
VS	O
Taiwan.exe	O


	O
libcef.dll	O


	O
Trojan	O
.	O
Win32.TONEINS	O


	O
~$20220817.docx	O


	O
~$20220617(1).docx	O


	O
Description	O

	O
First	O
-	O
stage	O

	O
legitimate	O

	O
executable	O
for	O

	O
DLL	O

	O
sideloading	O

	O
First	O
-	O
stage	O

	O
malware	O

	O
Second	O
-	O
stage	O

	O
legitimate	O

	O
executable	O
for	O

	O
DLL	O
sideloading	O


	O
Backdoor	O
.	O
Win32.TONESHELL	O


	O
Second	O
-	O
stage	O

	O
malware	O


	O
11/30	O


	O
15-8-2022.docx	O


	O
Decoy	O

	O
document	O


	O
China	O
VS	O

	O
Taiwan(1).docx	O


	O
Decoy	O

	O
document	O


	O
Table	O
3	O
.	O
Files	O
in	O
the	O
archive	O
of	O
Type	O
C	O

	O
libcef.dll	O
(	O
detected	O
by	O
Trend	O
Micro	O
as	O
Trojan	O
.	O
Win32.TONEINS	O
)	O
is	O
an	O
installer	O
for	O
the	O
nextstage	O
malware	O
.	O
It	O
copies	O
two	O
files	O
with	O
names	O
starting	O
with	O
"	O
~	O
"	O
,	O
in	O
this	O
case	O
,	O

	O
~$20220817.docx	O
and	O
~$20220617(1).docx	O
to	O
<	O
%	O
USERPROFILE%\Pictures	O
>	O
.	O
Both	O
files	O

	O
have	O
fake	O
file	O
extensions	O
and	O
masquerade	O
as	O
the	O
temporary	O
files	O
generated	O
while	O
opening	O

	O
Microsoft	O
Office	O
software	O
.	O


	O
Figure	O
9	O
.	O
Type	O
C	O
’s	O
malicious	O
routine	O

	O
Malware	O

	O
In	O
this	O
campaign	O
,	O
we	O
identified	O
the	O
following	O
malware	O
used	O
,	O
namely	O
PUBLOAD	O
,	O
TONEINS	O
,	O

	O
and	O
TONESHELL	O
.	O

	O
Trojan	O
.	O
Win32.PUBLOAD	O


	O
12/30	O


	O
PUBLOAD	O
is	O
a	O
stager	O
that	O
can	O
download	O
the	O
next	O
-	O
stage	O
payload	O
from	O
its	O
command	O
-	O
andcontrol	O
(	O
C&C	O
)	O
server	O
.	O
This	O
malware	O
was	O
first	O
disclosed	O
by	O
Cisco	O
Talos	O
in	O
May	O
2022	O
.	O

	O
Once	O
the	O
.dll	O
is	O
executed	O
,	O
it	O
first	O
checks	O
if	O
the	O
same	O
process	O
is	O
already	O
running	O
by	O
calling	O

	O
OpenEventA.	O
According	O
to	O
the	O
tweet	O
posted	O
by	O
Barberousse	O
,	O
some	O
noteworthy	O
event	O
names	O

	O
are	O
identified	O
as	O
usernames	O
of	O
other	O
cybersecurity	O
researchers	O
on	O
Twitter	O
,	O
such	O
as	O

	O
"	O
moto_sato	O
"	O
,	O
"	O
xaacrazyman_armyCIAx	O
,	O
"	O
and	O
"	O
JohnHammondTeam	O
.	O
"	O
It	O
is	O
important	O
to	O
note	O

	O
that	O
these	O
researchers	O
have	O
nothing	O
to	O
do	O
with	O
PUBLOAD	O
but	O
were	O
simply	O
and	O
intentionally	O

	O
mentioned	O
by	O
the	O
threat	O
actors	O
in	O
the	O
binaries	O
.	O


	O
Figure	O
10	O
.	O
An	O
example	O
of	O
the	O
special	O


	O
event	O
name	O
in	O
PUBLOAD	O

	O
Persistence	O

	O
PUBLOAD	O
creates	O
a	O
directory	O
in	O
<	O
C:\Users\Public\Libraries\	O
>	O
and	O
drops	O
all	O
the	O
malware	O
,	O

	O
including	O
the	O
malicious	O
DLL	O
and	O
the	O
legitimate	O
executable	O
,	O
into	O
the	O
directory	O
.	O
It	O
then	O
tries	O
to	O

	O
establish	O
persistence	O
in	O
one	O
of	O
the	O
following	O
ways	O
:	O

	O
1	O
.	O


	O
Adding	O
a	O
registry	O
run	O
key	O


	O
cmd.exe	O
/C	O
reg	O
add	O
HKCU\\Software\\Microsoft\\Windows\\CurrentVersion\\Run	O
/v	O

	O
Graphics	O
/t	O
REG_SZ	O
/d	O
\"Rundll32.exe	O
SHELL32.DLL	O
,	O
ShellExec_RunDLL	O

	O
\"C:\\Users\\Public\\Libraries\\Graphics\\AdobeLicensing.exe\"\	O
"	O
/f	O

	O
2	O
.	O


	O
Creating	O
a	O
schedule	O
task	O


	O
schtasks.exe	O
/F	O
/Create	O
/TN	O
Microsoft_Licensing	O
/sc	O
minute	O
/MO	O
1	O
/TR	O

	O
C:\\Users\\Public\\Libraries\\Graphics\\AdobeLicensing.exe	O

	O
Anti	O
-	O
Antivirus	O
:	O
API	O
with	O
callback	O


	O
13/30	O


	O
PUBLOAD	O
malware	O
decrypts	O
the	O
shellcode	O
in	O
AES	O
algorithm	O
in	O
memory	O
.	O
The	O
shellcode	O
is	O

	O
invoked	O
by	O
creating	O
a	O
thread	O
or	O
using	O
different	O
APIs	O
.	O
The	O
APIs	O
can	O
accept	O
an	O
argument	O
of	O
a	O

	O
callback	O
function	O
,	O
working	O
as	O
an	O
alternative	O
to	O
trigger	O
the	O
shellcode	O
.	O
We	O
observed	O
several	O

	O
leveraged	O
APIs	O
including	O
GrayStringW	O
,	O
EnumDateFormatsA	O
,	O
and	O
LineDDA	O
,	O
and	O
can	O
be	O

	O
considered	O
as	O
a	O
technique	O
to	O
bypass	O
antivirus	O
monitoring	O
and	O
detection	O
.	O


	O
Figure	O
11	O
.	O
An	O
example	O
of	O
shellcode	O
callback	O
in	O
PUBLOAD	O


	O
14/30	O


	O
Figure	O
12	O
.	O
APIs	O
that	O
accept	O
a	O
callback	O
function	O

	O
C&C	O
protocol	O

	O
The	O
decrypted	O
PUBLOAD	O
shellcode	O
collects	O
the	O
computer	O
name	O
and	O
the	O
username	O
as	O
the	O

	O
payload	O
of	O
the	O
first	O
beacon	O
.	O
The	O
payload	O
will	O
then	O
be	O
encrypted	O
with	O
the	O
predefined	O
RC4	O

	O
(	O
Rivest	O
Cipher	O
4	O
)	O
key	O
.	O
As	O
of	O
this	O
writing	O
,	O
all	O
the	O
stagers	O
we	O
have	O
seen	O
so	O
far	O
share	O
the	O
same	O

	O
key	O
.	O

	O
After	O
the	O
encryption	O
,	O
the	O
stager	O
uses	O
a	O
specific	O
byte	O
sequence	O
as	O
its	O
packet	O
’s	O
header	O
.	O
It	O

	O
prepends	O
the	O
magic	O
bytes	O
"	O
17	O
03	O
03	O
"	O
and	O
the	O
payload	O
size	O
before	O
the	O
encrypted	O
data	O
.	O


	O
Figure	O
13	O
.	O
The	O
RC4	O
key	O
used	O
(	O
top	O
)	O
and	O
the	O
packet	O
body	O
in	O
PUBLOAD	O
malware	O
(	O
bottom	O
)	O

	O
Name	O


	O
Offset	O


	O
Size	O


	O
Description	O

	O
15/30	O


	O
magic	O


	O
0x0	O


	O
0X3	O


	O
17	O
03	O
03	O


	O
size	O


	O
0x3	O


	O
0x2	O


	O
Payload	O
size	O


	O
payload	O


	O
0x5	O


	O
[	O
size	O
]	O


	O
Payload	O


	O
Table	O
4	O
.	O
Request	O
packet	O
format	O
in	O
PUBLOAD	O

	O
The	O
stager	O
also	O
checks	O
if	O
the	O
response	O
packet	O
has	O
the	O
same	O
magic	O
header	O
,	O
“	O
17	O
03	O
03	O
”	O
.	O
If	O
so	O
,	O

	O
the	O
downloaded	O
payload	O
in	O
memory	O
will	O
be	O
treated	O
as	O
a	O
piece	O
of	O
shellcode	O
and	O
will	O
be	O

	O
executed	O
directly	O
.	O

	O
Noteworthy	O
debug	O
strings	O

	O
In	O
early	O
2022	O
,	O
we	O
found	O
some	O
samples	O
of	O
PUBLOAD	O
embedded	O
with	O
debug	O
strings	O
.	O
They	O
are	O

	O
used	O
to	O
distract	O
analysts	O
from	O
the	O
main	O
infection	O
routines	O
.	O


	O
Figure	O


	O
14	O
.	O
The	O
distracting	O
debug	O
strings	O
in	O
PUBLOAD	O

	O
After	O
US	O
House	O
Speaker	O
Nancy	O
Pelosi	O
’s	O
visit	O
to	O
Taiwan	O
in	O
August	O
,	O
we	O
found	O
an	O
archive	O
file	O

	O
named	O
"	O
裴洛西訪台後民意匯總.rar	O
"	O
(	O
translated	O
as	O
“	O
The	O
public	O
opinion	O
summary	O
of	O
Pelosi	O
's	O

	O
visit	O
to	O
Taiwan	O
”	O
)	O
in	O
Traditional	O
Chinese	O
,	O
but	O
we	O
could	O
only	O
get	O
one	O
of	O
the	O
malicious	O
DLLs	O

	O
inside	O
the	O
archive	O
file	O
.	O
Since	O
the	O
topic	O
indicated	O
in	O
the	O
file	O
name	O
itself	O
is	O
considered	O
a	O

	O
controversial	O
topic	O
,	O
it	O
appears	O
potentially	O
catchy	O
to	O
the	O
targeted	O
recipient	O
.	O
The	O
DLL	O
turned	O
out	O

	O
to	O
be	O
a	O
PUBLOAD	O
stager	O
with	O
several	O
output	O
debug	O
strings	O
.	O


	O
16/30	O


	O
17/30	O


	O
Figure	O
15	O
.	O
Debug	O
strings	O
in	O
PUBLOAD	O

	O
Trojan	O
.	O
Win32.TONEINS	O

	O
Trojan	O
.	O
Win32.TONEINS	O
is	O
the	O
installer	O
for	O
TONESHELL	O
backdoors	O
.	O
The	O
installer	O
drops	O
the	O

	O
TONESHELL	O
malware	O
to	O
the	O
%	O
PUBLIC%	O
folder	O
and	O
establishes	O
the	O
persistence	O
for	O
it	O
.	O

	O
TONEINS	O
malware	O
usually	O
comes	O
in	O
the	O
lure	O
archives	O
,	O
and	O
in	O
most	O
cases	O
,	O
the	O
name	O
of	O
the	O

	O
TONEINS	O
DLL	O
is	O
libcef.dll	O
.	O
The	O
malicious	O
routine	O
is	O
triggered	O
via	O
calling	O
its	O
export	O
function	O

	O
cef_api_hash	O
.	O

	O
The	O
TONEINS	O
malware	O
is	O
obfuscated	O
,	O
likely	O
to	O
slow	O
down	O
malware	O
analysis	O
.	O
It	O
contains	O
a	O
lot	O

	O
of	O
junk	O
codes	O
in	O
its	O
control	O
flow	O
and	O
has	O
plenty	O
of	O
useless	O
XOR	O
instructions	O
as	O
though	O
to	O

	O
imply	O
that	O
these	O
are	O
used	O
to	O
decode	O
strings	O
.	O
Upon	O
checking	O
,	O
we	O
found	O
that	O
these	O
obfuscated	O

	O
codes	O
were	O
reused	O
from	O
an	O
open	O
-	O
source	O
repository	O
.	O


	O
Figure	O
16	O
.	O
Code	O


	O
obfuscation	O
in	O
TONEINS	O

	O
The	O
installer	O
establishes	O
the	O
persistence	O
for	O
TONESHELL	O
backdoors	O
by	O
using	O
the	O
following	O

	O
schtasks	O
command	O
:	O

	O
schtasks	O
/create	O
/sc	O
minute	O
/mo	O
2	O
/tn	O
"	O
ServiceHub	O
.	O
TestWindowStoreHost	O
"	O
/tr	O

	O
"	O
C:\Users\Public\Pictures\ServiceHub	O
.	O
TestWindowStoreHost.exe	O
"	O
/f	O

	O
Based	O
on	O
our	O
observations	O
,	O
the	O
file	O
names	O
for	O
the	O
dropped	O
TONESHELL	O
malware	O
differ	O
in	O

	O
case	O
,	O
and	O
so	O
do	O
the	O
names	O
of	O
the	O
scheduled	O
tasks	O
.	O
After	O
persistence	O
is	O
established	O
,	O

	O
TONESHELL	O
then	O
copies	O
the	O
legitimate	O
executable	O
and	O
the	O
malicious	O
DLL	O
to	O
the	O

	O
%	O
PUBLIC%	O
folder	O
,	O
wherein	O
both	O
files	O
have	O
names	O
that	O
start	O
with	O
“	O
~	O
”	O
in	O
the	O
lure	O
archive	O
.	O
In	O

	O
this	O
sample	O
,	O
~$20220817.docx	O
is	O
a	O
legitimate	O
executable	O
used	O
for	O
DLL	O
sideloading	O
,	O
and	O

	O
~$20220617(1).docx	O
is	O
the	O
TONESHELL	O
backdoor	O
DLL	O
to	O
be	O
installed	O
.	O


	O
18/30	O


	O
Figure	O
17	O
.	O
Files	O
with	O
fake	O
file	O
extensions	O

	O
Backdoor	O
.	O
Win32.TONESHELL	O

	O
The	O
TONESHELL	O
malware	O
is	O
the	O
main	O
backdoor	O
used	O
in	O
this	O
campaign	O
.	O
It	O
is	O
a	O
shellcode	O

	O
loader	O
that	O
loads	O
and	O
decodes	O
the	O
backdoor	O
shellcode	O
with	O
a	O
32	O
-	O
byte	O
key	O
in	O
memory	O
.	O
In	O
the	O

	O
earlier	O
version	O
of	O
TONESHELL	O
,	O
it	O
has	O
the	O
capabilities	O
from	O
TONEINS	O
malware	O
,	O
including	O

	O
establishing	O
persistence	O
and	O
installing	O
backdoors	O
.	O
However	O
,	O
the	O
more	O
recent	O
version	O
of	O

	O
TONESHELL	O
is	O
a	O
standalone	O
backdoor	O
without	O
any	O
installer	O
capabilities	O
(	O
such	O
as	O
the	O
file	O

	O
~$Talk	O
points.docx	O
)	O
.	O
It	O
is	O
also	O
obfuscated	O
in	O
a	O
similar	O
fashion	O
to	O
TONEINS	O
malware	O
,	O

	O
indicating	O
that	O
the	O
actors	O
continue	O
to	O
update	O
the	O
arsenal	O
and	O
separate	O
the	O
tools	O
in	O
order	O
to	O

	O
bypass	O
detection	O
.	O

	O
Anti	O
-	O
Analysis	O
:	O
Process	O
name	O
check	O

	O
In	O
order	O
to	O
make	O
sure	O
that	O
the	O
TONESHELL	O
is	O
installed	O
correctly	O
,	O

	O
Backdoor	O
.	O
Win32.TONESHELL	O
first	O
checks	O
if	O
the	O
process	O
path	O
matches	O
the	O
expected	O
one	O
.	O
If	O

	O
so	O
,	O
the	O
malicious	O
code	O
could	O
be	O
triggered	O
by	O
the	O
custom	O
exception	O
handler	O
.	O

	O
Figure	O
18	O
.	O
Process	O
name	O
check	O
in	O
TONESHELL	O

	O
Anti	O
-	O
Analysis	O
:	O
Custom	O
exception	O
handler	O
in	O
C++	O

	O
Interestingly	O
,	O
the	O
adversary	O
hides	O
the	O
actual	O
code	O
flow	O
with	O
the	O
implementation	O
of	O
custom	O

	O
exception	O
handlers	O
.	O
Different	O
exception	O
handlers	O
will	O
be	O
invoked	O
based	O
on	O
the	O
result	O
of	O
the	O

	O
process	O
name	O
check	O
,	O
continuing	O
the	O
malicious	O
routine	O
by	O
triggering	O
the	O
exception	O
with	O
the	O

	O
call	O
_	O
CxxThrowException	O
.	O
After	O
it	O
is	O
invoked	O
,	O
the	O
C++	O
runtime	O
will	O
find	O
the	O
corresponding	O

	O
exception	O
handler	O
from	O
the	O
ThrowInfo	O
structure	O
all	O
the	O
way	O
down	O
to	O
the	O
CatchProc	O
member	O
in	O

	O
the	O
_	O
msRttiDscr	O
structure	O
,	O
which	O
contains	O
the	O
real	O
malicious	O
codes	O
.	O
In	O
this	O
sample	O
,	O
the	O

	O
exception	O
handler	O
is	O
located	O
at	O
the	O
offset	O
0x10005300	O
.	O
This	O
technique	O
not	O
only	O
hides	O
the	O

	O
execution	O
flow	O
but	O
also	O
stops	O
the	O
execution	O
of	O
the	O
analyst	O
's	O
debugger	O
.	O


	O
19/30	O


	O
Figure	O
19	O
.	O
Data	O
workflow	O
of	O
exception	O
handling	O
in	O
C++	O
;	O
the	O
CatchProc	O
member	O
in	O
the	O
yellow	O

	O
circle	O
is	O
the	O
malicious	O
exception	O
handler	O
to	O
be	O
invoked	O


	O
Figure	O
20	O
.	O
The	O
main	O
malicious	O
routine	O
in	O
the	O
exception	O
handler	O

	O
Anti	O
-	O
Analysis	O
:	O
ForegroundWindow	O
check	O

	O
Looking	O
at	O
more	O
recent	O
TONESHELL	O
samples	O
,	O
we	O
noticed	O
that	O
a	O
new	O
anti	O
-	O
sandbox	O

	O
technique	O
is	O
added	O
compared	O
to	O
the	O
earlier	O
versions	O
.	O
The	O
newer	O
versions	O
invoke	O
the	O

	O
GetForegroundWindow	O
API	O
twice	O
and	O
check	O
if	O
there	O
is	O
any	O
window	O
switch	O
.	O
If	O
the	O
environment	O

	O
is	O
a	O
sandbox	O
,	O
both	O
calls	O
will	O
get	O
the	O
same	O
window	O
handle	O
because	O
there	O
is	O
no	O
human	O

	O
interaction	O
involved	O
in	O
most	O
sandboxes	O
,	O
resulting	O
in	O
the	O
foreground	O
window	O
not	O
changing	O
.	O
In	O

	O
addition	O
,	O
as	O
an	O
anti	O
-	O
sandbox	O
and	O
delayed	O
execution	O
technique	O
,	O
the	O
malicious	O
routine	O
can	O
only	O

	O
be	O
triggered	O
if	O
the	O
foreground	O
window	O
has	O
already	O
been	O
switched	O
for	O
the	O
fifth	O
time	O
.	O


	O
Figure	O


	O
21	O
.	O
GetForegroundWindow	O
check	O
in	O
newer	O
TONESHELL	O
samples	O

	O
20/30	O


	O
Figure	O
22	O
.	O
Malicious	O
routine	O
triggered	O
on	O
the	O
fifth	O
window	O
switch	O

	O
Shellcode	O
decoding	O

	O
After	O
the	O
malicious	O
exception	O
handler	O
is	O
triggered	O
,	O
it	O
starts	O
to	O
decode	O
the	O
next	O
-	O
stage	O

	O
TONESHELL	O
shellcode	O
.	O
To	O
decode	O
the	O
shellcode	O
,	O
it	O
first	O
decodes	O
a	O
32	O
-	O
byte	O
key	O
in	O
XOR	O

	O
operations	O
with	O
0x7D	O
,	O
and	O
the	O
key	O
will	O
then	O
be	O
used	O
to	O
decode	O
the	O
shellcode	O
body	O
.	O


	O
Figure	O
23	O
.	O
An	O


	O
example	O
of	O
the	O
32	O
-	O
byte	O
key	O
(	O
top	O
)	O
and	O
the	O
TONESHELL	O
shellcode	O
before	O
decoding	O
(	O
middle	O
)	O

	O
and	O
after	O
decoding	O
(	O
bottom	O
)	O

	O
Evolving	O
variants	O

	O
After	O
our	O
analysis	O
and	O
further	O
threat	O
hunting	O
,	O
we	O
found	O
several	O
variants	O
of	O
TONESHELL	O

	O
shellcode	O
:	O

	O
First	O
observed	O


	O
Variant	O


	O
Protocol	O


	O
C&C	O
encryption	O


	O
Supported	O
functions	O


	O
May	O
2022	O


	O
A	O


	O
Raw	O
TCP	O


	O
RC4	O


	O
File	O
upload	O

	O
File	O
download	O

	O
File	O
execution	O

	O
Lateral	O
movement	O


	O
Jul	O
2022	O


	O
B	O


	O
Raw	O
TCP	O


	O
32	O
-	O
byte	O
XOR	O


	O
File	O
upload	O

	O
Lateral	O
movement	O


	O
Sep	O
2021	O


	O
C	O


	O
HTTP	O


	O
RC4	O


	O
File	O
upload	O

	O
File	O
execution	O


	O
Table	O
5	O
.	O
Differences	O
between	O
TONESHELL	O
variants	O

	O
Variant	O
A	O


	O
21/30	O


	O
TONESHELL	O
supports	O
up	O
to	O
10	O
C&C	O
servers	O
by	O
design	O
,	O
but	O
in	O
all	O
the	O
samples	O
we	O

	O
encountered	O
only	O
one	O
C&C	O
server	O
was	O
used	O
.	O
Before	O
connecting	O
to	O
the	O
C&C	O
server	O
,	O
it	O

	O
generates	O
a	O
victim	O
ID	O
(	O
the	O
variable	O
unique_id	O
)	O
with	O
the	O
victim	O
's	O
volume	O
serial	O
and	O
the	O

	O
computer	O
name	O
,	O
or	O
with	O
a	O
randomly	O
generated	O
GUID	O
.	O


	O
Figure	O
24	O
.	O
Finding	O
10	O
C&C	O
servers	O
supported	O
in	O
TONESHELL	O


	O
Figure	O
25	O
.	O
The	O
algorithm	O
used	O
to	O
generate	O
the	O
victim	O
’s	O
ID	O
in	O
TONESHELL	O
variant	O
A	O

	O
In	O
the	O
first	O
beacon	O
,	O
it	O
collects	O
the	O
following	O
data	O
from	O
the	O
victim	O
's	O
machine	O
and	O
sends	O
them	O
to	O

	O
the	O
C&C	O
server	O
:	O

	O
1	O
.	O
Current	O
process	O
ID	O

	O
2	O
.	O
Volume	O
serial	O

	O
3	O
.	O
Username	O

	O
4	O
.	O
Computer	O
name	O

	O
5	O
.	O
Product	O
name	O

	O
6	O
.	O
Operating	O
system	O
bit	O

	O
7	O
.	O
Processes	O
list	O

	O
TONESHELL	O
communicates	O
over	O
raw	O
TCP	O
,	O
with	O
the	O
request	O
header	O
and	O
the	O
response	O

	O
header	O
starting	O
with	O
the	O
specific	O
magic	O
byte	O
sequence	O
"	O
17	O
03	O
03	O
"	O
.	O
Based	O
on	O
our	O
research	O
,	O

	O
this	O
magic	O
header	O
is	O
used	O
in	O
all	O
TONESHELL	O
TCP	O
variants	O
and	O
the	O
identified	O
PUBLOAD	O

	O
22/30	O


	O
malware	O
.	O
The	O
payload	O
in	O
the	O
packet	O
will	O
be	O
encrypted	O
in	O
RC4	O
algorithm	O
.	O
In	O
this	O
variant	O
,	O
its	O

	O
request	O
packet	O
format	O
is	O
as	O
follows	O
:	O

	O
Name	O


	O
Offset	O


	O
Size	O


	O
Description	O


	O
magic	O


	O
0x0	O


	O
0x3	O


	O
17	O
03	O
03	O


	O
size	O


	O
0x3	O


	O
0x2	O


	O
Payload	O
size	O


	O
type	O


	O
0x5	O


	O
0x1	O


	O
Connection	O
type	O
,	O
0x0	O
or	O
0x1	O


	O
unique_id	O


	O
0x6	O


	O
0x4	O


	O
Victim	O
ID	O


	O
payload	O


	O
0x10	O


	O
[	O
size	O
]	O


	O
Payload	O


	O
Table	O
6	O
.	O
Request	O
packet	O
format	O
in	O
TONESHELL	O
variant	O
A	O


	O
Figure	O
26	O
.	O
Packet	O
header	O
check	O
in	O

	O
TONESHELL	O
(	O
all	O
TCP	O
variants	O
and	O
stagers	O
)	O

	O
The	O
backdoor	O
supports	O
various	O
functions	O
,	O
including	O
file	O
upload	O
,	O
file	O
download	O
,	O
file	O
execution	O
,	O

	O
and	O
lateral	O
movement	O
.	O
We	O
also	O
noticed	O
that	O
its	O
internal	O
strings	O
are	O
self	O
-	O
explanatory	O
.	O
In	O
fact	O
,	O

	O
this	O
malware	O
is	O
named	O
TONESHELL	O
after	O
the	O
typo	O
found	O
in	O
its	O
command	O
"	O
TOnePipeShell	O
"	O
.	O

	O
The	O
following	O
table	O
shows	O
all	O
of	O
its	O
commands	O
:	O

	O
Code	O


	O
Internal	O
string	O


	O
Additional	O
description	O


	O
0x1	O


	O
-	O


	O
Reset	O
OnePipeShell	O
&	O
TwoPipeShell	O


	O
0x7	O


	O
-	O


	O
Reset	O
OnePipeShell	O
&	O
TwoPipeShell	O


	O
0x3	O


	O
-	O


	O
Unknown	O


	O
0x4	O


	O
-	O


	O
Change	O
sleep	O
seconds	O


	O
0x1A	O


	O
Upload	O
file	O

	O
begin	O


	O
0x1B	O


	O
Upload	O
file	O

	O
begin	O


	O
0x1D	O


	O
Upload	O
file	O

	O
cancel	O


	O
0x1C	O


	O
Upload	O
file	O

	O
Endup	O


	O
0x10	O


	O
Exec	O
file	O

	O
23/30	O


	O
0x21	O


	O
Create	O

	O
TOnePipeShell	O


	O
0x22	O


	O
OnePipeShell	O

	O
Close	O


	O
0x1E	O


	O
TwoPipeShell	O

	O
Create	O


	O
0x1F	O


	O
TwoPipeShell	O

	O
Write	O
File	O


	O
0x20	O


	O
TwoPipeShell	O

	O
Close	O


	O
0x18	O


	O
Download	O


	O
0x19	O


	O
CDownUpLoad	O


	O
0x21	O


	O
-	O


	O
OnePipeShell	O
:	O
one	O
-	O
way	O
shell	O
over	O
one	O
named	O
pipe	O
(	O
meant	O
for	O

	O
data	O
exchange	O
on	O
intranet	O
)	O


	O
TwoPipeShell	O
:	O
two	O
-	O
way	O
shell	O
over	O
two	O
named	O
pipes	O
(	O
meant	O
for	O

	O
data	O
exchange	O
on	O
intranet	O
)	O


	O
Exit	O


	O
Table	O
7	O
.	O
Command	O
codes	O
in	O
TONESHELL	O
variant	O
A	O

	O
Variant	O
B	O

	O
TONESHELL	O
variant	O
B	O
is	O
slightly	O
different	O
from	O
variant	O
A	O
wherein	O
the	O
victim	O
ID	O
is	O
generated	O

	O
from	O
the	O
tick	O
count	O
,	O
username	O
,	O
and	O
computer	O
name	O
instead	O
.	O


	O
Figure	O
27	O
.	O
Different	O


	O
algorithm	O
for	O
the	O
victim	O
ID	O
generation	O
in	O
TONESHELL	O
variant	O
B	O

	O
The	O
backdoor	O
's	O
protocol	O
is	O
also	O
different	O
.	O
The	O
payload	O
in	O
the	O
packet	O
is	O
encoded	O
with	O
a	O

	O
random	O
32	O
-	O
byte	O
key	O
,	O
and	O
the	O
key	O
differs	O
from	O
packet	O
to	O
packet	O
.	O
The	O
new	O
key	O
is	O
generated	O

	O
whenever	O
a	O
new	O
request	O
is	O
made	O
.	O

	O
Name	O


	O
Offset	O


	O
Size	O


	O
Description	O


	O
magic	O


	O
0x0	O


	O
0x3	O


	O
17	O
03	O
03	O


	O
size	O


	O
0x3	O


	O
0x2	O


	O
Payload	O
size	O


	O
24/30	O


	O
key	O


	O
0x5	O


	O
0x20	O


	O
32	O
-	O
byte	O
XOR	O
key	O


	O
payload	O


	O
0x25	O


	O
[	O
size	O
]	O


	O
Payload	O


	O
Table	O
8	O
.	O
Request	O
packet	O
format	O
in	O
TONESHELL	O
variant	O
B	O


	O
Figure	O


	O
28	O
.	O
In	O
TONESHELL	O
variant	O
B	O
,	O
the	O
payload	O
will	O
be	O
encoded	O
in	O
XOR	O
operations	O
before	O
a	O

	O
request	O
is	O
made	O
.	O

	O
The	O
command	O
codes	O
in	O
this	O
variant	O
are	O
as	O
follows	O
:	O

	O
Code	O


	O
Internal	O
string	O


	O
Description	O


	O
0x9	O


	O
-	O


	O
Reset	O
OnePipeShell	O


	O
0xA	O


	O
-	O


	O
Reset	O
OnePipeShell	O


	O
0x3	O


	O
-	O


	O
Unknown	O


	O
0x4	O


	O
-	O


	O
Change	O
sleep	O
seconds	O


	O
0x4	O


	O
Upload	O
file	O
begin	O


	O
0x5	O


	O
Upload	O
file	O
write	O


	O
0x7	O


	O
Upload	O
file	O
cancel	O


	O
0x6	O


	O
Upload	O
file	O
Endup	O


	O
0x3	O


	O
Create	O
TOnePipeShell	O


	O
Table	O
9	O
.	O
Command	O
codes	O
in	O
TONESHELL	O
Variant	O
B	O

	O
Variant	O
C	O

	O
During	O
our	O
research	O
,	O
we	O
hunted	O
a	O
dumped	O
TONESHELL	O
shellcode	O
from	O
VirusTotal	O
(	O
SHA256	O
:	O

	O
521662079c1473adb59f2d7134c8c1d76841f2a0f9b9e6e181aa54df25715a09	O
)	O
.	O
Our	O
analysis	O

	O
showed	O
it	O
works	O
similar	O
to	O
the	O
two	O
different	O
variants	O
,	O
but	O
the	O
C&C	O
protocol	O
used	O
is	O
HTTP	O
.	O
It	O

	O
25/30	O


	O
seems	O
to	O
be	O
the	O
earlier	O
version	O
of	O
TONESHELL	O
because	O
the	O
sample	O
was	O
uploaded	O
in	O

	O
September	O
2021	O
,	O
and	O
uses	O
the	O
POST	O
method	O
for	O
the	O
first	O
beacon	O
.	O
The	O
following	O
data	O
is	O

	O
collected	O
from	O
the	O
victim	O
's	O
machine	O
:	O

	O
1	O
.	O
Memory	O
size	O

	O
2	O
.	O
Username	O

	O
3	O
.	O
Computer	O
name	O

	O
4	O
.	O
Disk	O
size	O

	O
5	O
.	O
Operating	O
system	O
bit	O

	O
6	O
.	O
Product	O
name	O


	O
Figure	O
29	O
.	O


	O
The	O
first	O
HTTP	O
beacon	O
request	O
in	O
TONESHELL	O
Variant	O
C	O

	O
The	O
victim	O
's	O
ID	O
(	O
specified	O
by	O
the	O
"	O
Guid	O
"	O
header	O
in	O
the	O
first	O
beacon	O
and	O
later	O
used	O
in	O
the	O

	O
"	O
Cookie	O
"	O
header	O
)	O
is	O
also	O
generated	O
from	O
a	O
random	O
GUID	O
.	O
The	O
body	O
is	O
also	O
encrypted	O
in	O
RC4	O
,	O

	O
and	O
the	O
command	O
codes	O
are	O
much	O
like	O
Variant	O
B	O
as	O
follows	O
:	O

	O
Code	O


	O
Internal	O
string	O


	O
Additional	O
description	O


	O
0x2	O


	O
-	O


	O
Reset	O
OnePipeShell	O


	O
0x7	O


	O
-	O


	O
Reset	O
OnePipeShell	O


	O
0x3	O


	O
-	O


	O
Unknown	O


	O
0x4	O


	O
-	O


	O
Change	O
sleep	O
seconds	O


	O
0x1A	O


	O
Upload	O
file	O
begin	O


	O
0x1B	O


	O
Upload	O
file	O
write	O


	O
26/30	O


	O
0x1D	O


	O
Upload	O
file	O
cancel	O


	O
0x1C	O


	O
Upload	O
file	O
Endup	O


	O
0x10	O


	O
Exec	O
file	O


	O
Table	O
10	O
.	O
Command	O
codes	O
in	O
TONESHELL	O
variant	O
C	O


	O
Threat	O
hunting	O

	O
We	O
observed	O
that	O
several	O
TONESHELL	O
and	O
TONEINS	O
malware	O
samples	O
were	O
uploaded	O
to	O

	O
VirusTotal	O
in	O
recent	O
months	O
.	O
With	O
the	O
help	O
of	O
these	O
,	O
we	O
collected	O
several	O
Google	O
Drive	O
links	O
,	O

	O
such	O
as	O
770d5b60d8dc0f32941a6b530c9598df92a7ec76b60309aa8648f9b3a3f3cca5	O
.	O


	O
Figure	O
30	O
.	O
Example	O
of	O
a	O
Google	O
Drive	O
link	O
,	O
found	O
in	O
the	O
wild	O
,	O
containing	O
both	O
TONESHELL	O

	O
and	O
TONEINS	O

	O
Usually	O
,	O
we	O
see	O
such	O
download	O
links	O
as	O
the	O
first	O
arrival	O
vectors	O
.	O
The	O
Google	O
Drive	O
direct	O

	O
download	O
link	O
is	O
represented	O
in	O
the	O
format	O
https[:]//drive.google.com	O
/	O
uc	O
?	O

	O
id	O
=	O
gdrive_file_id&export	O
=	O
download	O
.	O
The	O
gdrive_file_id	O
is	O
a	O
unique	O
identifier	O
for	O
this	O
specific	O

	O
file	O
.	O
We	O
can	O
switch	O
to	O
web	O
viewer	O
to	O
check	O
its	O
file	O
contents	O
and	O
its	O
owner	O
by	O
modifying	O
the	O

	O
URL	O
:	O
https[:]//drive.google.com	O
/	O
file	O
/	O
d	O
/	O
gdrive_file_id	O
/	O
view	O
.	O

	O
In	O
the	O
details	O
panel	O
,	O
we	O
can	O
find	O
the	O
owner	O
of	O
this	O
file	O
,	O
and	O
by	O
hovering	O
on	O
the	O
icon	O
we	O
can	O
get	O

	O
the	O
email	O
address	O
.	O


	O
Figure	O
31	O
.	O
The	O
web	O
viewer	O
of	O
Google	O
Drive	O


	O
27/30	O


	O
Figure	O
32	O
.	O
The	O
file	O
owner	O
's	O


	O
name	O
and	O
email	O
address	O

	O
We	O
can	O
conduct	O
further	O
research	O
with	O
this	O
specific	O
email	O
account	O
.	O
For	O
example	O
,	O
after	O
our	O

	O
investigation	O
,	O
we	O
know	O
that	O
the	O
actors	O
abused	O
the	O
same	O
email	O
address	O
to	O
store	O
the	O
lure	O

	O
archives	O
in	O
Google	O
Drive	O
,	O
as	O
well	O
as	O
deliver	O
the	O
phishing	O
email	O
.	O
If	O
we	O
hunt	O
for	O
this	O
specific	O

	O
email	O
address	O
in	O
the	O
monitoring	O
logs	O
,	O
we	O
might	O
find	O
more	O
distributed	O
malware	O
.	O


	O
Attribution	O

	O
The	O
observed	O
TTPs	O
in	O
this	O
campaign	O
are	O
similar	O
to	O
the	O
campaign	O
mentioned	O
by	O

	O
Secureworks	O
.	O
Both	O
campaigns	O
abused	O
the	O
.lnk	O
files	O
to	O
trigger	O
the	O
malware	O
.	O
Compared	O
to	O
the	O

	O
said	O
report	O
’s	O
observations	O
,	O
the	O
archive	O
we	O
found	O
in	O
this	O
campaign	O
share	O
similar	O
folder	O

	O
structures	O
.	O


	O
Figure	O
33	O
.	O
Similar	O
folder	O
structure	O
of	O
BRONZE	O
PRESIDENT	O
(	O
left	O
)	O
and	O
Earth	O
Preta	O
(	O
right	O
)	O

	O
Based	O
on	O
the	O
same	O
report	O
,	O
Bronze	O
President	O
was	O
known	O
to	O
be	O
leveraging	O
APIs	O
with	O
a	O

	O
callback	O
function	O
argument	O
to	O
invoke	O
the	O
shellcode	O
like	O
EnumThreadWindows	O
.	O
Similar	O

	O
techniques	O
are	O
also	O
used	O
in	O
PUBLOAD	O
malware	O
.	O

	O
In	O
addition	O
,	O
we	O
also	O
spotted	O
a	O
link	O
between	O
the	O
two	O
campaigns	O
:	O
One	O
of	O
the	O
C&C	O
servers	O

	O
(	O
98[.]142[.]251[.]29	O
)	O
can	O
be	O
correlated	O
to	O
a	O
shortcut	O
file	O
.	O
This	O
shortcut	O
file	O
appears	O
in	O
one	O
lure	O

	O
archive	O
“	O
EU	O
31st	O
session	O
of	O
the	O
Commission	O
on	O
Crime	O
Prevention	O
and	O
Criminal	O
Justice	O

	O
United	O
Nations	O
on	O
Drugs	O
and	O
Crime.rar	O
”	O
(	O
SHA256	O
:	O

	O
09fc8bf9e2980ebec1977a8023e8a2940e6adb5004f48d07ad34b71ebf35b877	O
)	O
,	O
which	O
the	O

	O
Secureworks	O
report	O
also	O
mentioned	O
.	O
We	O
used	O
the	O
tool	O
LECmd	O
to	O
parse	O
the	O
shortcut	O
files	O

	O
wherein	O
we	O
found	O
the	O
specific	O
C&C	O
string	O
inside	O
the	O
metadata	O
of	O
the	O
.lnk	O
file	O
.	O
It	O
seems	O
that	O

	O
the	O
actor	O
used	O
the	O
C&C	O
string	O
as	O
the	O
folder	O
name	O
.	O


	O
28/30	O


	O
Figure	O
34	O
.	O
Metadata	O
of	O
the	O
.lnk	O
file	O
(	O
SHA256	O
:	O

	O
a693b9f9ffc5f4900e094b1d1360f7e7b907c9c8680abfeace34e1a8e380f405	O
)	O

	O
Third	O
,	O
the	O
infection	O
chains	O
mentioned	O
by	O
Cisco	O
Talos	O
also	O
resemble	O
what	O
we	O
have	O
observed	O

	O
recently	O
:	O

	O
1	O
.	O
Both	O
use	O
schtasks	O
and	O
registry	O
run	O
key	O
for	O
persistence	O
.	O

	O
2	O
.	O
Both	O
use	O
benign	O
executables	O
for	O
DLL	O
sideloading	O
.	O

	O
3	O
.	O
Both	O
use	O
malicious	O
archives	O
for	O
arrival	O
vectors	O
.	O

	O
Most	O
importantly	O
,	O
the	O
stager	O
mentioned	O
in	O
the	O
report	O
uses	O
the	O
same	O
magic	O
header	O
(	O
17	O
03	O
03	O
)	O

	O
as	O
TONESHELL	O
does	O
in	O
the	O
C&C	O
communication	O
protocol	O
,	O
thereby	O
solidifying	O
these	O
malware	O

	O
families	O
’	O
link	O
to	O
Earth	O
Preta	O
.	O


	O
Conclusion	O

	O
Earth	O
Preta	O
is	O
a	O
cyberespionage	O
group	O
known	O
to	O
develop	O
their	O
own	O
loaders	O
in	O
combination	O

	O
with	O
existing	O
tools	O
like	O
PlugX	O
and	O
Cobalt	O
Strike	O
for	O
compromise	O
.	O
Recent	O
research	O
papers	O

	O
show	O
that	O
it	O
is	O
constantly	O
updating	O
its	O
toolsets	O
and	O
indicate	O
that	O
it	O
is	O
further	O
expanding	O
its	O

	O
capabilities	O
.	O

	O
Based	O
on	O
our	O
analysis	O
,	O
once	O
the	O
group	O
has	O
infiltrated	O
a	O
targeted	O
victim	O
’s	O
systems	O
,	O
the	O

	O
sensitive	O
documents	O
stolen	O
can	O
be	O
abused	O
as	O
the	O
entry	O
vectors	O
for	O
the	O
next	O
wave	O
of	O

	O
intrusions	O
.	O
This	O
strategy	O
largely	O
broadens	O
the	O
affected	O
scope	O
in	O
the	O
region	O
involved	O
.	O
For	O
the	O

	O
group	O
’s	O
objectives	O
,	O
the	O
targeted	O
area	O
appears	O
to	O
be	O
the	O
countries	O
in	O
Asia	O
.	O

	O
As	O
part	O
of	O
organizational	O
mitigation	O
plans	O
,	O
we	O
recommend	O
implementing	O
continuous	O
phishing	O

	O
awareness	O
trainings	O
for	O
partners	O
and	O
employees	O
.	O
We	O
advise	O
always	O
checking	O
the	O
sender	O
and	O

	O
the	O
subject	O
twice	O
before	O
opening	O
an	O
email	O
,	O
especially	O
with	O
an	O
unidentifiable	O
sender	O
or	O
an	O

	O
unknown	O
subject	O
.	O
We	O
also	O
recommend	O
a	O
multi	O
-	O
layered	O
protection	O
solution	O
is	O
recommended	O
to	O

	O
detect	O
and	O
block	O
threats	O
as	O
far	O
left	O
to	O
the	O
malware	O
infection	O
chain	O
as	O
possible	O
.	O

	O
MITRE	O
ATT&CK	O


	O
29/30	O


	O
MITRE	O
ATT&CK	O
table	O


	O
Indicators	O
of	O
Compromise	O
(	O
IOCs	O
)	O

	O
The	O
full	O
list	O
of	O
IOCs	O
can	O
be	O
found	O
here	O
.	O


	O
30/30	O