5 matches found
trust-my-agent-ai
🛡️ TrustMyAgent 面向 OpenClaw 智能体的安全保镖 网站 • 信任中心 • 快速开始 • 贡献指南 为 OpenClaw 智能体提供的类似EDR的安全智能体。运行 41 项安全检查,涵盖 14 个领域,计算信任分数(0-100),并将遥测数据发送到集中的 信任中心 仪表盘,人类和其他智能体可以在其中验证可信度。 无状态设计 - 完全在内存中运行,不在本地存储任何内容,不在主机上留下痕迹。 为什么 AI智能体功能强大,但不够透明。当一个智能体在机器上运行时,你如何知道它没有: 泄露环境变量中的机密? 产生可疑进程? 访问不应访问的文件? 以过高权限运行? 连接外泄服务?...
TrustworthyRAG
Trustworthy RAG(可信赖的 RAG) 一个用于检测检索增强生成(RAG)系统中错误信息与知识投毒的评估代理。 概述 本项目实现了一个 Trustworthy RAG 框架 ,其中集成有评估代理(Evaluation Agent) ,通过三个互补的分析组件评估 RAG 响应的可靠性: NLI 验证器(NLI Verifier) - 通过自然语言推断(BART-MNLI)进行事实一致性检查 投毒检测器(Poison Detector) - 多信号对抗性内容检测(语言学、结构、语义、文档内/跨文档 NLI) 信任指数计算器(Trust Index Calculator) -...
Memory Poisoning Attack and Defense on Memory Based LLM-Agents
Large language model agents equipped with persistent memory are vulnerable to memory poisoning attacks, where adversaries inject malicious instructions through query only interactions that corrupt the agents long term memory and influence future responses. Recent work demonstrated that the MINJA...
Securing GenAI Multi-Agent Systems against Tool Squatting: a Zero Trust Registry-Based Approach
The rise of generative AI GenAI multi-agent systems MAS necessitates standardized protocols enabling agents to discover and interact with external tools. However, these protocols introduce new security challenges, particularly; tool squatting; the deceptive registration or representation of tools...
Software company accused of illegally profiling millions of mobile phone users
A digital rights and privacy organization has filed a complaint against software company TeleSign for gathering and selling information on millions of mobile phone users. The organization that filed the complaint is nyob. nyob is an Austrian based digital right organization that focusses on...