Releases: Samsung/CredSweeper
Releases · Samsung/CredSweeper
v1.3.1
What's Changed
- MlValidator object is created once
- CredSweeper is installed with pip as CLI tool additionally
- CredSweeper may be used as github action
- Added size limit option to skip large files in scan
v1.3.0
v1.2.1
Change list:
- Limit max
tensorflow
version by<2.8.0
due to the newtensorflow
release that changes API
Fix issue with tensorflow:
$ python -m credsweeper --path tests/samples/password --ml_validation
Traceback (most recent call last):
File "/home/user/CredSweeper/credsweeper/ml_model/ml_validator.py", line 12, in <module>
from tensorflow.python.keras.preprocessing.sequence import pad_sequences
ModuleNotFoundError: No module named 'tensorflow.python.keras.preprocessing'
New tensorflow==2.8.0
might be supported in 1.2.2
, 1.3.0
or later versions
v1.2.0
Short description:
- Make execution time ~5x faster
- Less CLI output if result is saved to json
- Fix crash in case of non-utf default system encoding
- Add option to specify ML threshold (less or more sensitive)
- New field in the output: float "ml_probability" which is non empty if ml validation was used
Please note:
Faster execution partially based on changes to the credsweeper/rules/config.yaml
If you modified this file in your deployment - please apply similar modifications to a newer version!
v1.1.0
Change list:
- Add diff scan for
.patch
files. See--diff_path PATH
CLI argument - Add support for non UTF-8 encoding in files
- Add 5 new regex rules: Firebase Domain, AWS S3 Bucket, Nonce, Salt, Certificate
- New severity type: Info
- Update ML model. Retrain it with new rules (mentioned above) and updated dataset
- Add batching for ML model, so it would run in parallel (or even on GPU if
tensorflow-gpu
configured by the user) - Add value field for the output
Add fist release on pypi: https://pypi.org/project/credsweeper/