Project Sekai - ✅-reversing-gauntlet

Project Sekai

🔒 RITSEC CTF 2023 / ✅-reversing-gauntlet

Gauntlet - 500 points

Category: Reversing Description: Four parts of a key are being assembled within this file. The adversary has implemented four corresponding anti-debug checks that need to be bypassed. Good luck! Files:

https://ctf.ritsec.club/files/c57ed94c23073566aabb767a5d410799/gauntlet.exe?token=eyJ1c2VyX2lkIjo3MTcsInRlYW1faWQiOjQ1OCwiZmlsZV9pZCI6MTd9.ZCcD9w.lQPxCiQKjPIA4VNQiwQl_-WsPAU

Tags: No tags.

Sutx pinned a message to this channel. 03/31/2023 9:02 AM

@crazyman ai wants to collaborate

@Iyed wants to collaborate

@sahuang

10:41

any progress on this one?

no i havent started it

RS{Quick_check_vector_this_exception_important_time_is_hash_Wa-kcdftcteeioi}

12:26

I'm stuck on the last part

12:26

calculating a hash

ooh

it checks if we have set any breakpoints

12:26

loops over the ThirdChallenge function

12:27

checks for 0xcc/0xe9

12:27

and there is a counter that it crements

12:27

I don't know what that hash on the flag I sent is incorrect even tho I haven't set any breakpoints

12:28

and I tried to return an arbitrary value myself setting rax to the ThirdChallenge function length but still getting a wrong flag when submitting on the platform

Iyed

RS{Quick_check_vector_this_exception_important_time_is_hash_Wa-kcdftcteeioi}

wait is this 3 parts or all 4?

all 4

12:29

the last part uses the ThirdChallenge functoin

12:29

to calculate the hash

hm

12:29

last part seems not correct from the whole word sense?

I think it's just a hash so it shouldn't make sense

ic

12:30

let me check

Image attachment

what hash is _hash_?>

RS{Quick_check_vector_this_exception_important_time_is_hash_WihclehrtegckPiecidyptf.ntftoelttepeTrigolohde}

12:36

just a thing it calculates

12:36

from a string

12:37

it takes characters and concatenates them

12:38

Image attachment

12:38

only these 2 parts matter right? we shouldn't enter those 2 only ig

12:39

I tried to patch it with nops and only let the 0x90 but still wrong (edited)

what are first 3 parts flags?

12:39

splitted

RS{Quick_check_
vector_this_exception_
important_time_is
_hash_
hash
}

@TheBadGod wants to collaborate

BreakCheck just checks for breakpoints

12:42

because debuggers insert a 0xcc instruction, which can be detected like this

yes but haven't set anyones

12:45

and I patched that check

yeah but like there are only these possibilities for after the hash thingy:

b'Withinthischallenge,therearefouranti-debuggerchecks.Passingeachcheckwilldecryptpartoftheflag.Onlythefirsttwocheckswillnotifyyouiftheywerebypassed.Theremainingtwowillnot.Author:kindtime\x00'
b'Wtihscalnehrrorat-eugrcek.Psigechcildcytprftefa.nytefrttocekilntfofteeebpse.Termiigtowlo.uhridie'
b'Whicln,herfrn-bgesPsgaeldrtael.Oyhfstesityoityeps.hraioittrkdm'
b'WihclehrtegckPiecidyptf.ntftoelttepeTrigolohde'
b'WnsleerutbrcPnceietttlOtithfutwbsTeiwltu:d'
b'Wcnhrr-gePgdte.yfteitotep.riord'
b'Wta,u-eksckdfltscwtueeioti'
b'WhlhtgkiiyfnfotpTioohd'
b'Winerbeslt.Ofiyt.aotrm'
b'WsertrPcittttftbTilud'
b'Wfecgkycr.nlti'
b'WchrgPd.fettprod'
b'Whrncnfeiesilr\x00'
b'Wa-kcdftcteeioi'
b'WlrbPetlOttsitd'
b'Wlgiifpiod'
b'Weoeeenfrrli'
b'Wnret.fit.or'
b'Wgasctoybn\x00'
b'WetPitttTld'
b'W,-sdlstit'
b'Wegyc.li'
b'Wtuapywau'
b'Whgftpod'
b'WerettuTt'
b'Wrcfeeir'
b'Weel.Oiom'
b'Wkdtteo'
b'Wa.rioh'

right?

12:51

we can rule out the ones with a null byte as we still need to close the flag with } (edited)

so we try one by one?

gimme a sec

12:56

mmh, also getting Wa-kcdftcteeioi

12:56

(e.count(b"\xe9")*5+len(list(filter(lambda x: x!=0xe9 and x!=0x90, e)))-e.count(b"\x90"))%30 14

(edited)

13:01

but on check two im not getting the same

13:01

b'evector_this_exception_\x00'

yeah pretty sure it should be RS{Quick_check_vector_this_exception_important_time_is_hash_Wa-kcdftcteeioi}

13:22

actually getting one underscore more, but that's gotta be timing issues

Image attachment

13:26

yeah second time around I got the RS{Quick_check_vector_this_exception_important_time_is_hash_Wa-kcdftcteeioi}

"we're expecting something different after that W" from what admin told

ok, sadly i ran it without a debugger and only then attached the debugger, so they are kinda wrong about that

TheBadGod

ok, sadly i ran it without a debugger and only then attached the debugger, so they are kinda wrong about that

author said chall could be wrong so he will check it

but idk how 3 solves come from

13:38

kindtime — Today at 1:36 PM Hi all, Gauntlet author here. It’s come to my understanding that there might be an issue with the challenge. I won’t be able to check it for a few hours, but if you are having issues, please come back to it another time. I’ll throw a message in here once I have the opportunity. Thanks

13:38

lol

bruh

13:38

also 4 solves now

LOL wtf

like there can't be anything other than a W given the code we have

yeah they said after the W

13:40

but still

13:40

prob sth wrong with them lol

TheBadGod

yeah but like there are only these possibilities for after the hash thingy:

b'Withinthischallenge,therearefouranti-debuggerchecks.Passingeachcheckwilldecryptpartoftheflag.Onlythefirsttwocheckswillnotifyyouiftheywerebypassed.Theremainingtwowillnot.Author:kindtime\x00'
b'Wtihscalnehrrorat-eugrcek.Psigechcildcytprftefa.nytefrttocekilntfofteeebpse.Termiigtowlo.uhridie'
b'Whicln,herfrn-bgesPsgaeldrtael.Oyhfstesityoityeps.hraioittrkdm'
b'WihclehrtegckPiecidyptf.ntftoelttepeTrigolohde'
b'WnsleerutbrcPnceietttlOtithfutwbsTeiwltu:d'
b'Wcnhrr-gePgdte.yfteitotep.riord'
b'Wta,u-eksckdfltscwtueeioti'
b'WhlhtgkiiyfnfotpTioohd'
b'Winerbeslt.Ofiyt.aotrm'
b'WsertrPcittttftbTilud'
b'Wfecgkycr.nlti'
b'WchrgPd.fettprod'
b'Whrncnfeiesilr\x00'
b'Wa-kcdftcteeioi'
b'WlrbPetlOttsitd'
b'Wlgiifpiod'
b'Weoeeenfrrli'
b'Wnret.fit.or'
b'Wgasctoybn\x00'
b'WetPitttTld'
b'W,-sdlstit'
b'Wegyc.li'
b'Wtuapywau'
b'Whgftpod'
b'WerettuTt'
b'Wrcfeeir'
b'Weel.Oiom'
b'Wkdtteo'
b'Wa.rioh'

right?

well here are all the possibilities

TheBadGod

(e.count(b"\xe9")*5+len(list(filter(lambda x: x!=0xe9 and x!=0x90, e)))-e.count(b"\x90"))%30 14

(edited)

here how I calculate the amount that's returned

TheBadGod

we can rule out the ones with a null byte as we still need to close the flag with } (edited)

also ignore this, it actually checks for strlen, which i fucked up

13:44

WetPitttTld

13:44

they can't fucking count

13:44

just tried all of them

TheBadGod

used /ctf solve

✅ Challenge solved.

LOL wait what

they fucked up, ly3dMejri and I both got 14, it got 14 in the program without a debugger and attaching afterwards

oh you mean they counted the length wrong?

but they say it's like idk 25 or sth

sahuang

oh you mean they counted the length wrong?

they probably counted, then changed some code which means different amount of instructions => flag changes

ah

13:46

IC

because they just do -#0x90 bytes + 5*#0xe9 bytes + #all other bytes

13:46

mod 30

ok so hash changed when they prob changed sth and recompiled

yeah

13:47

probably

nice

15:48

I was out

15:49

gg

Exported 88 message(s)