oracle-samples/ml-model-dynscan
A security-focused framework for dynamically scanning and testing machine learning models in an isolated Docker environment, detecting suspicious behaviors such as unauthorized network access or system modifications.
This framework helps ML developers and security engineers detect suspicious behaviors in machine learning models before deployment. It takes an ML model and a custom test script as input, then runs the model in a secure, isolated Docker environment to identify unauthorized network access, file system modifications, or dangerous system calls. The output is a report detailing any potential security issues found, ensuring the integrity and safety of the models.
No commits in the last 6 months.
Use this if you need to automatically scan your machine learning models for malicious or risky behaviors in a controlled environment.
Not ideal if you're looking for a tool to evaluate model performance or accuracy rather than security vulnerabilities.
Stars
11
Forks
—
Language
Python
License
UPL-1.0
Category
Last pushed
Aug 05, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/oracle-samples/ml-model-dynscan"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
nmap/nmap
Nmap - the Network Mapper. Github mirror of official SVN repository.
e-m-b-a/emba
EMBA - The firmware security analyzer
ait-testbed/attackbed
The AttackBed is a simulated enterprise network with numerous vulnerabilities. Attacks in this...
ritesh-gupta-git/AI-Powered-Vulnerability-Management
AI-VMF: AI-Powered Vulnerability Management Framework demo (ensemble exploit prediction + risk scoring)
scorpiondefense/cyberweapons
Automated Cyber Offense