model-poisoning

A curated list of MLSecOps tools and resources for securing machine learning and AI systems - adversarial ML defense, LLM security, AI red teaming, model scanning, supply-chain protection, and MLOps pipeline security.

Comprehensive taxonomy of AI security vulnerabilities, LLM adversarial attacks, prompt injection techniques, and machine learning security research. Covers 71+ attack vectors including model poisoning, agentic AI exploits, and privacy breaches.

“AI model poisoning industry” (highlighted in the 315 consumer protection program) shows that manipulating AI outputs via crafted content is no longer theoretical.

WeightsWatcher is a cryptographic integrity verification system for Machine Learning artifacts

Federated learning + secure aggregation + zero-knowledge norm-bound defence

Train one machine-learning model across many parties who don't trust each other. A dishonest minority can't skew the result, and every party computes byte-for-byte identical output on any CPU or architecture (big-endian included) — so each round ships a receipt anyone can re-check offline and reproduce exactly. Rust; drop-in for Flower.

Reproducible TDAW research lab for history-aware Byzantine-robust federated MNIST training, with a Rust core, Flower modes, and a Next.js dashboard.

Learning repository for Foundations of AI Security by AttackIQ Security Academy. Includes notes, labs, reports, case studies, and certificate of completion, focusing on adversarial threats, defense strategies, and AI security testing.

Adversarial Model Update Poisoning (AMUP) threat model and Update Instability Score (UIS) for covert poisoning detection in AI supply chains. Accepted at IEEE Connect 2026.