Boyan Milanov

We’ve added a pickle file scanner to Fickling that uses an allowlist approach to protect AI/ML environments from malicious pickle files that could compromise models or infrastructure.

Datasig generates compact, unique fingerprints for AI/ML datasets that let you compare training data with high accuracy—without needing access to the raw data itself.
This critical capability helps AIBOM (AI bill of materials) tools detect data-borne vulnerabilities that traditional security tools completely miss.

In part 1, we introduced Sleepy Pickle, an attack that uses malicious pickle files to stealthily compromise ML models and carry out sophisticated attacks against end users. Here we show how this technique can be adapted to enable long-lasting presence on compromised systems while remaining undetected. This variant technique, which we call […]

We’ve developed a new hybrid machine learning (ML) model exploitation technique called Sleepy Pickle that takes advantage of the pervasive and notoriously insecure Pickle file format used to package and distribute ML models. Sleepy pickle goes beyond previous exploit techniques that target an organization’s systems when they deploy ML models to instead […]

We have released Maat, a cross-architecture, multi-purpose, and user-friendly symbolic execution framework. It provides common symbolic execution capabilities such as dynamic symbolic execution (DSE), taint analysis, binary instrumentation, environment simulation, and constraint solving. Maat is easy-to-use, is based on the popular Ghidra intermediate representation (IR) language p-code, prioritizes runtime performance, and has […]