Skip to content

GenAI-ML-SecAudit is an implementation of OWASP 2025 Top 10 ML and Gen AI Apps vulnerabilities test on Tafsiri-S2ST GenAI App. The scripts simulate attacks, capture logs, and generate an interactive HTML graph that visualizes the results.

License

Notifications You must be signed in to change notification settings

mbwika/GenAI-ML-SecAudit

Repository files navigation

GenAI-ML-SecAudit

It implements a simulation of the OWASP 2025 Top 10 ML and Gen AI Apps vulnerabilities attacks on Tafsiri (a speech-to-speech translator S2ST GenAI app). The scripts automate attacks, capture logs, and generate an interactive HTML graph that visualizes the results.

SETUP

  1. Change the API endpoint to match the actual endpoint your model is listening to.
  2. Change input parameters to match what your API endpoint expects (to avoid 422 error).

EXECUTE

  1. To test without an actual application, execute report_generator.py. It will use sample data provided.
  2. To run actual tests, execute run_pipeline.py
image newplot

About

GenAI-ML-SecAudit is an implementation of OWASP 2025 Top 10 ML and Gen AI Apps vulnerabilities test on Tafsiri-S2ST GenAI App. The scripts simulate attacks, capture logs, and generate an interactive HTML graph that visualizes the results.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published