**Show HN: ModelFuzz – Open-source runtime guardrails for AI agents**

> ModelFuzz secures LLM agents against prompt injection. Scan for vulnerabilities, then block unsafe tool calls at the execution layer with one decorator. The Threat Prompt injection turns your agent's own tools against you. A single poisoned document, email, or web page can hijack an LLM through indirect prompt injection, compromising LLM agent security at the source. The model thinks it's helping. It isn't. ! shell.r...

**来源信息**
- **来源**：Hacker News：AI 热帖
- **分类**：ai-products
- **发布时间**：2026-07-27 02:22（北京时间）
- **原文**：[打开原文](https://www.modelfuzz.com/)