cpggen 1.0.4

Generate CPG for multiple languages for use with joern

CPG Generator

 ██████╗██████╗  ██████╗
██╔════╝██╔══██╗██╔════╝
██║     ██████╔╝██║  ███╗
██║     ██╔═══╝ ██║   ██║
╚██████╗██║     ╚██████╔╝
 ╚═════╝╚═╝      ╚═════╝

CPG Generator is a python cli tool to generate Code Property Graph for multiple languages. The generated CPG can be directly imported to Joern or uploaded to Qwiet.AI for analysis.

Pre-requisites

• JDK 11 or above
• Python 3.10
• Docker or podman (Windows, Linux or Mac) or
• Joern natively installed (Linux only)

Installation

cpggen is available as a single executable binary, PyPI package or as a container image.

Single executable binaries

Download the executable binary for your operating system from the releases page. These binary bundle the following:

• Joern with all the CPG frontends
• cpggen with Python 3.10
• cdxgen with Node.js 18 - Generates SBoM

curl -LO https://github.com/AppThreat/cpggen/releases/download/v1.0.4/cpggen-linux-amd64
chmod +x cpggen-linux-amd64
./cpggen-linux-amd64 --help

On Windows,

curl -LO https://github.com/appthreat/cpggen/releases/download/v1.0.4/cpggen.exe
.\cpggen.exe --help

NOTE: On Windows, antivirus and antimalware could prevent this single executable from functioning properly. Depending on the system, administrative privileges might also be required. Use container-based execution as a fallback.

OCI Artifacts via ORAS cli

Use ORAS cli to download the cpggen binary on Linux and Windows.

VERSION="1.0.0"
curl -LO "https://github.com/oras-project/oras/releases/download/v${VERSION}/oras_${VERSION}_linux_amd64.tar.gz"
mkdir -p oras-install/
tar -zxf oras_${VERSION}_*.tar.gz -C oras-install/
sudo mv oras-install/oras /usr/local/bin/
rm -rf oras_${VERSION}_*.tar.gz oras-install/

oras pull ghcr.io/appthreat/cpggen-bin:v1
chmod +x cpggen-linux-amd64
./cpggen-linux-amd64 --help

On Windows

set VERSION="1.0.0"
curl.exe -sLO  "https://github.com/oras-project/oras/releases/download/v%VERSION%/oras_%VERSION%_windows_amd64.zip"
tar.exe -xvzf oras_%VERSION%_windows_amd64.zip
mkdir -p %USERPROFILE%\bin\
copy oras.exe %USERPROFILE%\bin\
set PATH=%USERPROFILE%\bin\;%PATH%

Invoke-WebRequest -Uri https://github.com/oras-project/oras/releases/download/v1.0.0/oras_1.0.0_windows_amd64.zip -UseBasicParsing -OutFile oras_1.0.0_windows_amd64.zip
Expand-Archive -Path oras_1.0.0_windows_amd64.zip -DestinationPath .
oras.exe pull ghcr.io/appthreat/cpggen-windows-bin:v1

PyPI package

This would install just the python cli tool without any CPG language frontends. Joern must be installed separately to make the cli work.

pip install cpggen

Bundled container image

docker pull ghcr.io/appthreat/cpggen
# podman pull ghcr.io/appthreat/cpggen

Almalinux 9 requires the CPU to support SSE4.2. For kvm64 VM use the Almalinux 8 version instead.

docker pull ghcr.io/appthreat/cpggen-alma8
# podman pull ghcr.io/appthreat/cpggen-alma8

Or use the nightly to always get the latest joern and tools.

docker pull ghcr.io/appthreat/cpggen:nightly
# podman pull ghcr.io/appthreat/cpggen:nightly

Usage

To auto detect the language from the current directory and generate CPG.

cpggen

To specify input and output directory.

cpggen -i <src directory> -o <CPG directory or file name>

You can even pass a git url as source

cpggen -i https://github.com/HooliCorp/vulnerable-aws-koa-app -o /tmp/cpg

To specify language type.

cpggen -i <src directory> -o <CPG directory or file name> -l java

# Comma separated values are accepted for multiple languages
cpggen -i <src directory> -o <CPG directory or file name> -l java,js,python

Container based invocation

docker run --rm -it -v /tmp:/tmp -v $(pwd):/app:rw --cpus=4 --memory=16g -t ghcr.io/appthreat/cpggen cpggen -i <src directory> -o <CPG directory or file name>

Export graphs

By passing --export, cpggen can export the various graphs to many formats using joern-export

Example to export cpg14 graphs in dot format

cpggen -i ~/work/sandbox/crAPI -o ~/work/sandbox/crAPI/cpg_out --build --export --export-out-dir ~/work/sandbox/crAPI/cpg_export

To export pdg in neo4jcsv format

cpggen -i ~/work/sandbox/crAPI -o ~/work/sandbox/crAPI/cpg_out --build --export --export-out-dir ~/work/sandbox/crAPI/cpg_export --export-repr pdg --export-format neo4jcsv

Slicing graphs

Pass --slice argument to extract intra-procedural slices from the CPG. By default, slices would be based on Usages. Pass --slice-mode DataFlow to create a sliced CPG based on DataFlow.

cpggen -i ~/work/sandbox/crAPI -o ~/work/sandbox/crAPI/cpg_out --slice

Artifacts produced

Upon successful completion, cpggen would produce the following artifacts in the directory specified under out_dir

• {name}-{lang}-cpg.bin.zip - Code Property Graph for the given language type
• {name}-{lang}-cpg.bom.xml - SBoM in CycloneDX XML format
• {name}-{lang}-cpg.bom.json - SBoM in CycloneDX json format
• {name}-{lang}-cpg.manifest.json - A json file listing the generated artifacts and the invocation commands

Server mode

cpggen can run in server mode.

cpggen --server

You can invoke the endpoint /cpg to generate CPG.

curl "http://127.0.0.1:7072/cpg?src=/Volumes/Work/sandbox/vulnerable-aws-koa-app&out_dir=/tmp/cpg_out&lang=js"

curl "http://127.0.0.1:7072/cpg?url=https://github.com/HooliCorp/vulnerable-aws-koa-app&out_dir=/tmp/cpg_out&lang=js"

Languages supported

Language	Requires build
C	No
C++	No
Java	No (*)
Scala	Yes
Jsp	Yes
Jar/War	No
JavaScript	No
TypeScript	No
Kotlin	No (*)
Php	No
Python	No
C# / dotnet	Yes
Go	Yes

(*) - Precision could be improved with dependencies

Environment variables

Name	Purpose
JOERN_HOME	Joern installation directory
CPGGEN_HOST	cpggen server host. Default 127.0.0.1
CPGGEN_PORT	cpggen server port. Default 7072
CPGGEN_CONTAINER_CPU	CPU units to use in container execution mode. Default computed
CPGGEN_CONTAINER_MEMORY	Memory units to use in container execution mode. Default computed
CPGGEN_MEMORY	Heap memory to use for frontends. Default computed
AT_DEBUG_MODE	Set to debug to enable debug logging
CPG_EXPORT	Set to true to export CPG graphs in dot format
CPG_EXPORT_REPR	Graph to export. Default all
CPG_EXPORT_FORMAT	Export format. Default dot
CPG_SLICE	Set to true to slice CPG
CPG_SLICE_MODE	Slice mode. Default Usages
SHIFTLEFT_ACCESS_TOKEN	Set to automatically submit the CPG for analysis by Qwiet AI

GitHub actions

Use the marketplace action to generate CPGs using GitHub actions. Optionally, the upload the generated CPGs as build artifacts use the below step.

- name: Upload cpg
  uses: actions/upload-artifact@v1.0.0
  with:
    name: cpg
    path: cpg_out

License

Apache-2.0

Developing / Contributing

git clone [e-mail]:AppThreat/cpggen.git
cd cpggen

python -m pip install --upgrade pip
python -m pip install poetry
# Add poetry to the PATH environment variable
poetry install

poetry run cpggen -i <src directory>