Cloak 帮你安全地把 API 密钥交给 AI 代理，模型和日志永远看不到密钥，防止提示注入泄露。
Article URL: https://github.com/cloakward/cloak Comments URL: https://news.ycombinator.com/item?id=48618904 Points: 2 # Comments: 0
LLM防护盾反成攻击靶心：揭示基于大模型的agent安全护栏存在拒绝服务新漏洞
arXiv:2606.14517v1 Announce Type: cross Abstract: LLM-based guardrails have emerged as a highly effective defense against prompt injection and jailbre…
10秒部署的本地代理，防止AI代理泄漏你的凭证，连它自己都记不住秘密。
A process can't leak a secret it never had. Shai-hulud, prompt-injection - you name it. They cannot steal what your agent (or an process) don't have. …