An AI Agent Execution Environment to Safeguard User Data

AI agents promise to serve as general-purpose personal assistants for their users, which requires them to have access to private user data e.g., personal and financial information. This poses a serious risk to security and privacy. Adversaries may attack the AI model e.g., via prompt injection to...

Applying Security Engineering to Prompt Injection Security

This seems like an important advance in LLM security against prompt injection: Google DeepMind has unveiled CaMeL CApabilities for MachinE Learning, a new approach to stopping prompt-injection attacks that abandons the failed strategy of having AI models police themselves. Instead, CaMeL treats...

IE GC information leakage the vulnerability of the gossip-vulnerability warning-the black bar safety net

This vulnerability is several months earlier dion cow release, also recently got a pwnie award of the prize, the original speaking of the flash, ff, etc. of the GC engine are the use of conserved marker removal algorithm and are not tag data or pointer, so the presence of this problem, dion cattl...