{"id":"seebug-paper-99f125318235","seq":1577348,"kind":"news","title":"基于小语言模型的 RAG 流水线上下文感知提示注入检测：nokast-secureRAG","url":"https://paper.seebug.org/3529","permalink":"https://vulnpulse.app/items/seebug-paper-99f125318235","summary":"作者：Abhishek Rai 原文链接：https://arxiv.org/html/2610.28364v1 摘要 本地检索增强生成（Retrieval-Augmented Generation, RAG）系统让强大的智能助手能够在私有数据上运行，但也继承了一类日益增长的安全风险——尤其是间接提示注入（Indirect Prompt Injection, IPI）与检索投毒，即隐藏在检索内容中...","severity":null,"techs":[],"cves":[],"tags":[],"source":{"id":"seebug-paper","name":"Seebug Paper (Knownsec 404)"},"published_at":"2026-10-23T08:33:28.000Z","fetched_at":"2026-10-27T09:57:34.907Z"}