Skip to content

h311d1n3r/Cerberus

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Cerberus

Description

A C++ tool to unstrip Rust and Go binaries (ELF and PE)

Cerberus is the tool you want to use to make RUST and GO static analysis a lot easier.
Based on hashing and scoring systems, it can retrieve lots of symbol names.

How does it work ?

After analyzing your ELF/PE binary to find the used libraries, Cerberus will download and build them.
Then the tool will hash (in various ways) the functions in your file and in the libraries to make matches.

Table of contents

Installation
      Download a release
      Build the tool with Docker
      Build the tool on host
How to use ?
      Syntax
      Parameters
      Flags
      Example
Warning

Check the Releases tab on the Github project and download the latest one.

  1. Clone the repository git clone https://github.com/h311d1n3r/Cerberus && cd cerberus.
  2. Check the available Dockerfiles under Cerberus/docker/{OS}.
  3. Build the docker image of your choice docker build -f ./docker/{OS}/Dockerfile-{version} ..
  4. You can run Cerberus from inside the docker or extract the binary on your host. This second choice needs to install the libraries listed in this section.
  1. You need to have libarchive, libcurl4-openssl, zlib1g, libelf and the uuid-dev libraries installed on your system.
    With APT just do apt -y install libarchive-dev libcurl4-openssl-dev zlib1g-dev libelf-dev
  2. Clone the repository git clone https://github.com/h311d1n3r/Cerberus && cd cerberus.
  3. Initialize git dependencies : git submodule update --init
  4. Create the build directory mkdir build && cd build.
  5. Run CMake to configure the project cmake ...
  6. Run make to compile the project make.

cerberus binary [-param value] [--flag]

output -> Specifies the path for the resulting ELF file.
part_hash_len -> Specifies the length of a part hash. The part hash of a function is just a reduction of the function with a linear pace. This technique is used to prevent fixed addresses from corrupting a standard hash. Default value : 20
part_hash_trust -> Specifies minimum ratio of similarity between the two hashed functions to compare. The kept function will be the one with the most matches anyway. Increasing this value will reduce the number of matched functions but speed up execution time. Default value : 0.6
min_func_size -> The minimum length a function must be to get analyzed. Decreasing this value will increase matches but also false positives. Default value : 10

help -> Displays a help message.
debug -> Displays outputs of commands.
no-prompt -> Automatically skips user prompts.

Command

The following command will try to unstrip the file ./rust_example into a new ELF called ./rust_example_syms.
cerberus ./rust_example -output ./rust_example_syms

Result

Here is a comparison of the main function in the two files using Binary Ninja :

before.png

after.png

This software must only be used to carry out lawful experiments and I am not responsible for any breach of this rule !