https://github.com/tsundokul/pyradamsa
Python bindings for calling radamsa mutators
https://github.com/tsundokul/pyradamsa
fuzzing security
Last synced: 5 months ago
JSON representation
Python bindings for calling radamsa mutators
- Host: GitHub
- URL: https://github.com/tsundokul/pyradamsa
- Owner: tsundokul
- License: mit
- Created: 2020-07-06T06:30:10.000Z (almost 6 years ago)
- Default Branch: master
- Last Pushed: 2021-10-07T18:57:56.000Z (over 4 years ago)
- Last Synced: 2025-09-03T13:25:11.161Z (10 months ago)
- Topics: fuzzing, security
- Language: Python
- Homepage:
- Size: 9.77 KB
- Stars: 26
- Watchers: 2
- Forks: 2
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
pyradamsa
==========================
`pyradamsa` provides an interface for calling libradamsa methods from within Python, allowing one to perform mutations on byte blocks (aka fuzzing). For more details see [radamsa](https://gitlab.com/akihe/radamsa) (a general-purpose fuzzer) and [libradamsa](https://github.com/andreafioraldi/libradamsa) (precompiled radamsa library).
### NOTE
The master branch uses AFL++'s libradamsa. An alternative version created by [p1-olm](https://github.com/p1-olm) is available on [radamsa-upstream](https://github.com/tsundokul/pyradamsa/tree/radamsa-upstream) branch and is based on the [original](https://gitlab.com/akihe/radamsa) libradamsa upstream.
How is the `radamsa-upstream` different than the `master` version?
- The owl VM initialization uses different seeds, so result are different when mutating same bytes with the same seed
- Future updates from libradamsa creators will show up first in `radamsa-upstream`
## Usage
Install the package using `pip` (or [build](#building) it from source)
```bash
pip install pyradamsa
```
Do some fuzzing
```python
import pyradamsa
rad = pyradamsa.Radamsa()
mydata = b'GET /auth?pass=HelloWorld HTTP1.1'
fuzzed = rad.fuzz(mydata, seed=1337)
print(fuzzed)
> b'GET /auth?pass=HelloWorld HTTP\xc0\xb1.1'
# seed is randomly set if not provided
rad.fuzz(mydata)
> b'\tG\xf3\xa0\x81\x9c\xf7dLET \xe2\x81/aut\xf3\xa0\x80\xafHTTP2.rld HTTP2.rld HTTP3.2\xe1\xa0\x8e9'
rad.fuzz(mydata)
> b'GET /auth?pass=HelloWorld HTTP1.340282366920938463463374607431768211455'
etc.
# enforce static seed on initialization
rad = pyradamsa.Radamsa(seed=0)
# max_mut enforces a maximum length for returned data
# it defaults to (data length + an offset of 4096 bytes)
fuzzed = rad.fuzz(mydata, seed=1337, max_mut=10)
> b'GET /auth?'
# the offset may be overwritten on init
rad = pyradamsa.Radamsa(mut_offset=2048)
```
## Building
Currently wheels are available for linux i686 and x86_64
```sh
# Clone the repo
git clone --recurse-submodules https://github.com/tsundokul/pyradamsa.git
cd pyradamsa
# patch memory leak when reinitializing owl vm
patch libradamsa/libradamsa.c realloc.patch
# OPTIONAL: when using manylinux (https://github.com/pypa/manylinux)
docker run --rm -it -v `pwd`:/io quay.io/pypa/manylinux2010_x86_64 /bin/bash
cd /io && alias python='/opt/python/cp35-cp35m/bin/python3.5'
export PATH="/opt/python/cp35-cp35m/bin/:${PATH}"
# Install requirements
python -m pip install -r requirements.txt
# Build C extension (libradamsa.so)
python setup.py build_ext
# Run tests
./run_tests
# Build wheel
python setup.py bdist_wheel
```
## Contributing
* Fork the repo
* Check out a feature or bug branch
* Add your changes
* Update README when needed
* Submit a pull request to upstream repo
* Add description of your changes
* Ensure tests are passing
* Ensure branch is mergeable
_MIT License, 2020_ [@tim17d](https://twitter.com/tim17d)