PyPI page
Home page
Author:
None
License:
MIT
Summary:
The full-stack safety layer for AI agents. 22 shields including prompt injection, toxicity, hallucination detection, data exfiltration, privilege escalation, and more — in 2 lines of code.
Latest version:
1.5.0
Optional dependencies:
anthropic
|
datasets
|
detoxify
|
furo
|
google-genai
|
llama-cpp-python
|
numpy
|
openai
|
pytest
|
pytest-asyncio
|
pyyaml
|
scikit-learn
|
sentence-transformers
|
sphinx
|
sphinx-copybutton
Downloads last day:
22
Downloads last week:
70
Downloads last month:
265