PyPI page
Home page
Author:
Vishnu Prashanth
License:
MIT
Summary:
Static checks for risky context-to-action flows in AI agents, MCP tools, and LLM applications. Detects patterns associated with prompt injection, tool poisoning, and unsafe agent memory/output handling. OWASP LLM Top 10 (2025) tagged.
Latest version:
0.5.1
Optional dependencies:
black
|
flake8
|
jsonschema
|
mypy
|
pytest
|
pytest-cov
Downloads last day:
0
Downloads last week:
4
Downloads last month:
274
7-day smoothing