The Board Room
guide to hacking
EFTA01735540
"GETYOURFACTSFIRST,
DISTORTTHEMASYOUPI:
MARKTWAIN
Lifehack Quotes .0
EFTA01735541
Are you compromised?
Yes
EFTA_R1_00019494
EFTA01735542
Why is everyone
compromised?
1 Your network is a replicable monoculture
2. Compromising is a one-way street: You can't "un-
compromise" something
3. The internet and your network are a graph of trust:
compromising is viral and exponential
4. Your defense is reactive and slow, it must be
proactive and fast
EFTA_R1_001,19495
EFTA01735543
fam-res
.... Ana - ;
••
•
•••.. ; - • • f••
- • • 'a
a
- • •
-• . • 41110 111-••
40. •
eifffir • •••• " • • -Mr
•
•
• .0, a.
• . a -t
•
r .‘44 •
a
• a
.
4. s . ea
.
-
all •a a a
a e _ .aa •
• • I••
IF.. • ti
a .1. f
4.- *a'
. ....
• a
_en=
r. • •
a -
aa.
OK IN
.41••
• • •••• •
•
• -
• • .
•
..t••
aft gp •
1. - • ' • •
•
•.• I
•
• •
. •
• fa. • ea
•
. alma
• r • ap
a a "re I iv •
,f
•
•
•
• „
••••
• • •
a
4 * •
• • •
I # I sob -
4
S
• 44
I
THAT'S YOUR
• I •
•
4
•
• • •
.
•
•
•
a. a
I • it •
-•
• •
ego
.
I
ok
. •
t t •
MPET I
•
• -
•
•
e w- • •t
OP• •
#
• •
• *4.
4
.• •
-•
• a .^ dr-
• 4 •
••••'"I "
•
1.
•
4
•
• •
•
•
„ • •••- caw'. - • •
• • p. L-
sea 41
V mite"r"
' •
••-a e'L • a
• • **- • 4 63•1*
•• • 49. _lb a i ii. • ;.
•
aS
eaSOO •
• -. 0 .....
•
•
• ... ....---- -
•
EFTA R1 00019496
EFTA01735544
Monoculture
• The attacker can download the same software you
have and attack it until he finds a way in.
• An attacker can replicate an almost-exact copy of
your machine and go at it until he finds an "in"
• Once the attacker is on a machine he can
experiment and explore the trusted neighbors until
he finds an "in"
EFTA_R1_00019497
EFTA01735545
EFTA R1 00019498
EFTA01735546
"Un-compromise"-able
• A maxim: there's always a deep enough level in a
machine that is not defended/defendable
• It used to be the kernel, now it's the bios, the
firmware, the hardware, the secret co-processor,
you name it
• You can't "un-compromise" because it's impossible
to know what's compromised
EFTA_R1_00019499
EFTA01735547
EFTA R1 00019500
EFTA01735548
Graphs of trust
• A lot of security today happens at the "perimeter",
once you're in it's game over. This is called "lateral
movement"
• Implicit trust: we trust somebody else servers to
download executables, we trust certificate
authorities keys, we trust our partner servers
• This means that your threat model is in large part
outside of your control
EFTA_R1_00019501
EFTA01735549
P:(201,114. :• io cr'IJC:
••
•
.1 •:•
Q• ;
• r •/ 1 ,
b
ie l e
-
0;
• 00 . A I-1 • *- • -. *--
-1 • ,- J
p•
.41•
.6
10 • 111•1,
• C II vi k
•
• -Pil e, • •..(cas•- • • - •
I
'..I • • •.;
S .- • -• • • '114 .0° • Le 11
/4• •
e i gOi •
A* .
re
(
ar/ t di ctith it
ar t
•• ••• :2. • ‘Vt ist
o,c.
r • • ! ••••'•-• • hit/ • ; . : r.•• ••••
• -••••••,....) ;•7411:.••••.; • • .
/,,,,.••tY „„ •
•
•„, .
.
.:•••.• • •1
•
•r ,
, -. . ,.. ^ . ...A•T 1,ii. • i •••• -4. of"..:':•.
• p • . . i I • , ••• r
P
• • •.- /1... ea. • i•1 '14 .1 fr.%
I ' •
....
• .•.•;
•
..
' .•
. .,
•
"roe,"
I I . ./•
d. •II•
1 it
•
" i f I . Ph
-
I. ••:', •••• .'..‘ Via vie
t •••: 7 , .iII/1"..A. I .. 1.... ••
" • • . ' ' i si •
r e•••••
•%•-.
•
I i•
.1 -1
f • • ' : •• PT •,•• i ••
•.siC VS l i' Ct ..•
1 • *). ••• e ." I *Al " %. l• • i f A 1')/
. y •
• • • •• : ] I. i l.te.•: . •5‘,..9 1. 1
1t i ( 1 . •
•, • • ‘ agt•ai •::$•4 1 1)4 4 • /..••
. • ; 11; . 4 I.'s'''. I
* .
• .•• r • •. •
. ..• • •
• ,.-k e'r
1
• ile4;
err.. • • • •Ir;•••;,.:'
•re • • 4
en) •• ••• r i" ' • • • , •• •• • I ; ' • ••• .*I •
. • ••. I •
EFTA01735550
Reactive and slow
• Most security tools today work by identifying an
attack somewhere else and then try to protect
everyone else
• This is reactive in nature and ineffective: most
attacks stay latent for a very long time
• Even with almost-real time detection, the attacker
needs to beat you at the race just once
EFTA_R1_00019503
EFTA01735551
EFTA_R1_00019504
EFTA01735552
The recursive guide to
compromise anything
1. Compromise a machine (exploit, social
engineering, backdoor, physical access)
2. The maxim: there's always a deep enough level in
a machine that is not defended/defendable. Go
there and stay put
3. For every node in the graph that trusts your
machine, go to 1 and be fast
EFTA_R1_0019505
EFTA01735553
Digital immune system
EFTA01735554
Digital immune system
• We have the technology to build 80% of the digital
immune system
• We need network effects and board-level decisions
to make the remaining 20% true
• This will not solve computer security but it will leap
it ahead by a lot
EFTA_R1_00019507
EFTA01735555
)
EFTA R1 00019508
EFTA01735556
"Shape-shifting" software
• No two copies of the same app, (kernel, firmware,
etc etc) should behave the same way at the micro
level
• Code should adapt to its users/owners, detect and
log anomalous behavior on a distributed ledger
EFTA_R1_00019509
EFTA01735557
"Accountability
breeds response-
ability."
-Stephen Covey
QuoteAddicts
EFTA01735558
Code Signing
• Every piece of code that is executed on a machine
should be signed by a trusted entity
• We can't trust a single company/machine: create a
distributed ledger of valid signatures for every
piece of code
EFTA_R1_00019511
EFTA01735559
I
EFTA_R1_000,19512
EFTA01735560
Self-destructing machines
• Every machine should have a "known-good" state
to revert to
• Every time a machine is thought to be
compromised it should be destroyed immediately
and reverted back to the "known-good" state
EFTA_R1_001,19513
EFTA01735561
EFTA_R1_00019514
EFTA01735562
Adaptive network structure
• The trusting neighbors of a machine must be able
to shut down communication with the allegedly
compromised machine
• The trusting neighbors should be able to adapt
their network topology to use a mirror copy of the
compromised machine
EFTA_R1_00019515
EFTA01735563
EFTA R1 00019516
EFTA01735564
The Al future
• In the future a lot of offensive security will be Al/ML-
driven
• In the future security will be much faster and much
more complicated
• We can't have proper defense against that without
these building blocks
EFTA_R1_001,19517
EFTA01735565
O&A
EFTA_R1_00019518
EFTA01735566