Skip to content

quarkslab/pyrrha

Repository files navigation

Pyrrha: A mapper collection for firmware analysis

Pyrrha is a filesystem cartography and correlation software focusing on visualization. It currently focuses on the relationship between executable files but aims at enabling anyone to map and visualize any relationship types. It uses the open-source code source explorer Sourcetrail to provide users with an easy way to navigate through and search for path to function.

An example of the symbols and libraries imported by libgcc_s.so.1 and of the symbols which reference this library.

An example of the symlinks which point on busybox.

Installation

The installation is done in two parts:

  • Installing Pyrrha as a Python module (pip install pyrrha-mapper) or using its Docker image.
  • Installing Sourcetrail to be able to visualize Pyrrha's results. You can use its last release and its documentation.

Usage

The usage workflow is composed of two steps which allow you to separate DB creation and result visualization.

  1. Run Pyrrha to obtain Sourcetrail compatible files (*.srctrlprj for the project file and *.srctrldb for the DB file). With the python package, you can just launch the command:
    $ pyrrha
    Usage: pyrrha [OPTIONS] COMMAND [ARGS]...
    
      Mapper collection for firmware analysis.
    
    Options:
      -h, --help  Show this message and exit.
    
    Commands:
      fs  Map PE and ELF files of a filesystem into a sourcetrail-compatible db.
    
    
    or with the Docker
    $ docker run  --rm -t -v $PWD:/tmp/pyrrha ghcr.io/quarkslab/pyrrha:latest [OPTIONS] COMMAND [ARGS]...
    
  2. Visualize your results with Sourcetrail
    $ sourcetrail PROJECT_NAME.srctrlprj
    

The detailed documentation of each mapper is available in the documentation.

Publications

Pyrrha has been presented at two conferences listed below. These talks include live demo of the fs parser which map links between libraries and executables files.

Authors

  • Eloïse Brocas (@ebrocas), Quarkslab

Past Contributors

  • Pascal Wu (@pwu42), during his internship at Quarkslab