Skip to content

Latest commit

 

History

History
241 lines (142 loc) · 8.82 KB

FEATURES.md

File metadata and controls

241 lines (142 loc) · 8.82 KB

emp3r0r

linux post-exploitation framework made by linux user


ARCHIVED, this file won't be updated, please check README


table of contents

what does it do

core features

transports

emp3r0r utilizes HTTP2 (TLS enabled) for its CC communication, but you can also encapsulate it in other transports such as TOR, and CDNs. all you need to do is tell emp3r0r agent to use your proxy

also, emp3r0r has its own CA pool, agents trusts only emp3r0r's own CA (which you can generate using build.py), making MITM attack much harder

below is a screenshot of emp3r0r's CC server, which has 3 agent coming from 3 different transports

ls_targets

auto proxy for agents without direct internet access

emp3r0r agents check if they have internet access on start, and start a socks5 proxy if they do, then they broadcast their proxy addresses (in encrypted form) on each network they can reach

if an agent doesn't have internet, its going to listen for such broadcasts. when it receives a working proxy, it starts a port mapping of that proxy and broadcasts it to its own networks, bringing the proxy to every agent it can ever touch, and eventually bring all agents to our CC server.

in the following example, we have 3 agents, among which only one ([1]) has internet access, and [0] has to use the proxy passed by [2]

autoproxy

anti-antivirus (or anti-whateveryoucallthem)

  • a cryptor that loads agent into memory
  • shellcode dropper
  • everything is randomized
  • one agent build for each target

agent traffic

every time an agent starts, it checks a preset URL for CC status, if it knows CC is offline, no further action will be executed, it waits for CC to go online

you can set the URL to a GitHub page or other less suspicious sites, your agents will poll that URL every random minutes

no CC communication will happen when the agent thinks CC is offline

if it isnt:

bare HTTP2 traffic:

traffic

when using Cloudflare CDN as CC frontend:

cdn

packer - start agent in memory

packer encrypts agent binary, and runs it from memory (using memfd_create)

currently emp3r0r is mostly memory-based, if used with this packer

packer

dropper - pure memory based agent launching

dropper drops a shellcode or script on your target, eventually runs your agent, in a stealth way

below is a screenshot of a python based shellcode delivery to agent execution:

dropper

hide processes and files

currently emp3r0r uses libemp3r0r to hide its files and processes, which utilizes glibc hijacking

persistence

currently implemented methods:

more will be added in the future

modules

reverse proxy

think it as ssh -R, when autoproxy module doesn't work because of the firewall on the agent that provides proxy service, what can you do?

in normal circumstances, we would use ssh -R to map our client-side port to the ssh server, so the server can connect to us to share our internet connection.

thats exactly what emp3r0r does, except it doesn't require any openssh binaries to be installed, type use reverse_proxy to get started!

with this feature you can bring every host that you can reach to emp3r0r CC server.

reverse_proxy

shellcode injection

inject guardian shellcode into arbitrary process, to gain persistence

shellcode injection

shellcode loader

this module helps you execute meterpreter or Cobalt Strike shellcode directly in emp3r0r's memory, combined with reverse_portfwd, you can use other post-exploitation frameworks right inside emp3r0r

shellcode loader

basic command shell

this is not a shell, it just executes any commands you send with sh -c and sends the result back to you

besides, it provides several useful helpers:

  • file management: put and get
  • command autocompletion
  • #net shows basic network info, such as ip a, ip r, ip neigh
  • #kill processes, and a simple #ps
  • bash !!! this is the real bash shell, keep on reading!

cmd shell

ssh to any kind of shells you like!

with module interactive_shell, you can set shell to normal bash, sh, busybox, or even python if you like!

all the shells works like you ssh to the host, for most cases, PTY is fully enabled

this is choosing a shell to ssh into, by default we are doing bash

ssh-shell

you can see the bash shell you just created in a new tmux window

bash

and python? you can spaw('bash') if you like

python

you can open as many shells as you like!

each shell has its own port mapping, allowing you to ssh -p port localhost directly.

with tmux you can see all of your shells organized cleanly in your current tmux session

shells

credential harvesting

not implemented yet

i wrote about this in my blog

auto root

currently emp3r0r supports CVE-2018-14665, agents can exploit this vulnerability if possible, and restart itself with root privilege

get_root.png

LPE suggest

upload the latest:

and run them on target system, return the results

lpe_suggest.png

port mapping

map any target addresses to CC side, using HTTP2 (or whatever transport your agent uses)

port_fwd.png

reverse port mapping (interoperability with other frameworks)

this screenshot shows a meterpreter session established with the help of emp3r0r

reverse port mapping

plugin system

yes, there is a plugin system. please read the wiki for more information

plugins.png

plugins-bash.png

thanks