---
id: 1494
title: "Exploiting ML models with pickle file attacks: Part 2"
source_domain: blog.trailofbits.com
source_url: "https://blog.trailofbits.com/2024/06/11/exploiting-ml-models-with-pickle-file-attacks-part-2/"
published_at: "2024-06-11 15:00"
tags: ["supply-chain", "malware", "python", "machine-learning", "pickle", "model-security"]
phying_url: "https://phying.com/item/1494"
lang: en
---

# Exploiting ML models with pickle file attacks: Part 2

## Summary

Part 2 of Trail of Bits' pickle attack series introduces Sticky Pickle: a self-replicating payload that hooks pickle.dump() to spread into re-saved models and obfuscates code to evade scanners.

## Intelligence Metadata

- **Sources:** blog.trailofbits.com
- **Published:** 2024-06-11 15:00
- **Original:** https://blog.trailofbits.com/2024/06/11/exploiting-ml-models-with-pickle-file-attacks-part-2/
- **Phying News:** https://phying.com/item/1494