This project is an AI-powered APK analysis tool that leverages the CodeGemma language model to assist in reverse engineering and deobfuscation of Android applications. It automates the process of decompiling APK files, extracting and analyzing Java code, and uses advanced natural language processing to identify potential flags, secrets.
1
stars
3
commits
Python
primary language
Aug 10, 2025
updated
An AI-powered APK analysis tool that leverages the CodeGemma language model to assist in reverse engineering and deobfuscation of Android applications.
pip install -r requirements.txt
.env file with your Hugging Face token:HUGGING_FACE_TOKEN=your_token_here
apk directorypython scripts/decompile.py # Decompile APK
python scripts/analyze.py # Run AI analysis
scripts/ - Python source code
decompile.py - APK decompilation logicai_model.py - CodeGemma integrationextract_flags.py - Flag detectionapk/ - APK files for analysisdecompiled/ - Decompiled Java codeai_output/ - Analysis results3 commits
Python
79.8%
JavaScript
13.4%
HTML
5.3%
CSS
1.4%
This project is an AI-powered APK analysis tool that leverages the CodeGemma language model to assist in reverse engineering and deobfuscation of Android applications. It automates the process of decompiling APK files, extracting and analyzing Java code, and uses advanced natural language processing to identify potential flags, secrets.
1
stars
3
commits
Python
primary language
Aug 10, 2025
updated
An AI-powered APK analysis tool that leverages the CodeGemma language model to assist in reverse engineering and deobfuscation of Android applications.
pip install -r requirements.txt
.env file with your Hugging Face token:HUGGING_FACE_TOKEN=your_token_here
apk directorypython scripts/decompile.py # Decompile APK
python scripts/analyze.py # Run AI analysis
scripts/ - Python source code
decompile.py - APK decompilation logicai_model.py - CodeGemma integrationextract_flags.py - Flag detectionapk/ - APK files for analysisdecompiled/ - Decompiled Java codeai_output/ - Analysis results3 commits
Python
79.8%
JavaScript
13.4%
HTML
5.3%
CSS
1.4%