RiccardoBiosas/awesome-MLSecOps

The collection organizes resources across multiple ML security dimensions including adversarial robustness, LLM vulnerability scanning, data privacy auditing, model serialization safety, and MLOps infrastructure vulnerabilities. It catalogs both practical attack frameworks (adversarial example generators, prompt injection tools) and defensive solutions (privacy-preserving libraries, model validation platforms) spanning TensorFlow, PyTorch, and other major ML ecosystems. Beyond tools, it includes threat matrices, CTF challenges, and best-practice documentation for securing ML pipelines end-to-end from training through production deployment.

425 stars. No commits in the last 6 months.