Knowledge-Driven Multi-Turn Jailbreaking on Large Language Models

Large Language Models LLMs face a significant threat from multi-turn jailbreak attacks, where adversaries progressively steer conversations to elicit harmful outputs. However, the practical effectiveness of existing attacks is undermined by several critical limitations: they struggle to maintain ...

Code Injection in archivy/archivy

Description Archivy is a self-hosted knowledge repository that allows you to safely preserve useful content that contributes to your knowledge bank. Vulnerability description Vulnerable to YAML deserialization attack caused by unsafe loading. Proof of Concept Run exploit.py python import os...