**Interlock: A runtime firewall for AI agents that assumes injection won**

> On building an exfiltration detector, the 46.7% false positive rate that forced a rewrite, and why I publish a list of things my own tool cannot catch. 31 July 2026 Interlock: a runtime firewall for AI agents that assumes prompt injection already won On building an exfiltration detector, the 46.7% false positive rate that forced a rewrite, and why I publish a list of things my own tool cannot catch. Here's a sentence...

**来源信息**
- **来源**：Hacker News：AI 热帖
- **分类**：ai-products
- **发布时间**：2026-08-06 01:11（北京时间）
- **原文**：[打开原文](https://yashwanthreddymali.com/blog/interlock-exfiltration-at-runtime)