AI Sentry - Unified AI/LLM Security Scanner for Static Code Analysis + Live Model Testing
Project description
aisentry
A unified command-line tool for AI/LLM security scanning. Static code analysis for OWASP LLM Top 10 vulnerabilities.
Website: aisentry.co | Docs: aisentry.co/docs
Benchmarks
Evaluated against a comprehensive OWASP LLM Top 10 testbed with 73 ground-truth vulnerabilities.
| Metric | aisentry | Semgrep | Bandit |
|---|---|---|---|
| Precision | 75.4% | 83.3% | 58.3% |
| Recall | 63.0% | 6.8% | 38.4% |
| F1 Score | 68.7% | 12.7% | 46.3% |
LLM-Specific Coverage (patterns generic tools miss):
| Category | aisentry | Semgrep | Bandit |
|---|---|---|---|
| LLM01: Prompt Injection | 72.7% | 0% | 15.4% |
| LLM04: Model DoS | 80.0% | 0% | 0% |
| LLM06: Sensitive Info | 62.5% | 0% | 0% |
| LLM10: Model Theft | 44.4% | 0% | 0% |
See docs/tool_comparison.md for detailed comparison and llm-sec-eval for methodology.
Features
- Static Code Analysis: OWASP LLM Top 10 + SQL injection detection
- Security Posture Audit: 61 controls across 10 categories with maturity scoring
- Remote Scanning: GitHub, GitLab, Bitbucket URLs
- Multiple Outputs: Text, JSON, HTML (interactive), SARIF (CI/CD)
- False Positive Reduction: ML-trained heuristics (88% accuracy)
Installation
pip install aisentry
# With ML-based false positive reduction
pip install aisentry[ml]
# With all cloud providers
pip install aisentry[cloud]
Quick Start
# Scan local project
aisentry scan ./my_project
# Scan GitHub repository
aisentry scan https://github.com/langchain-ai/langchain
# Generate HTML report
aisentry scan ./my_project -o html -f report.html
# Security posture audit
aisentry audit ./my_project
Live Model Testing
For runtime testing of LLM models (prompt injection, jailbreaks), we recommend Garak by NVIDIA. aisentry focuses on static code analysis - finding vulnerabilities before deployment.
Documentation
| Topic | Link |
|---|---|
| CLI Reference | docs/cli.md |
| Configuration | docs/configuration.md |
| Architecture | docs/architecture.md |
| CI/CD Integration | docs/integration.md |
| Tool Comparison | docs/tool_comparison.md |
GitHub Actions
- run: pip install aisentry
- run: aisentry scan . -o sarif -f results.sarif
- uses: github/codeql-action/upload-sarif@v2
with:
sarif_file: results.sarif
See docs/integration.md for GitLab, Azure DevOps, and pre-commit examples.
Development
git clone https://github.com/deosha/aisentry.git
cd aisentry
pip install -e ".[dev]"
pytest tests/ -v
License
MIT License - see LICENSE for details.
Links
- Website: aisentry.co
- GitHub: github.com/deosha/aisentry
- PyPI: pypi.org/project/aisentry
- Issues: Report bugs
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file aisentry-1.0.0.tar.gz.
File metadata
- Download URL: aisentry-1.0.0.tar.gz
- Upload date:
- Size: 750.2 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.13.6
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
3d31395289e5c2b3fae7c7222a8a57f2453b86ba0727194a8051c4b6fc6d7abd
|
|
| MD5 |
e71159a7e788e39a3e5ad843ba5dca3d
|
|
| BLAKE2b-256 |
de4129c068592b6fb184c375f8b253be1cc02884bc93f9014cd0ec9dcde6d678
|
File details
Details for the file aisentry-1.0.0-py3-none-any.whl.
File metadata
- Download URL: aisentry-1.0.0-py3-none-any.whl
- Upload date:
- Size: 486.0 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.13.6
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
ce644c7715bb3f1599348a2dfebe113c153bdc827b19f261f6db52ba41ea232e
|
|
| MD5 |
68804143e56d1322ba734bd6e5032511
|
|
| BLAKE2b-256 |
6abe2e998568d76b733aeb53a0c1d1d70caa2768fc9f8957e21f6b2e7a7e4315
|