# Brother Sefirot introduces Prompt-Firewall, a deterministic security plugin for AI agents

> Original title: Introducing Prompt-Firewall plugin.🪽

- Source: X search: AI launch posts (introducing, just launched)
- Published: 2026-11-11T03:39:26.000Z
- Bad Influence: https://badinfluence.news/items/dt8y1hilhzri9es18f5ewz370
- Original: https://x.com/SefirotWatch/status/2120360727525425206

## Summary

Brother Sefirot introduces Prompt-Firewall, a plugin that detects prompt injections in untrusted tool output and blocks dangerous actions for AI agents. The plugin runs locally with no LLM calls and no telemetry. It was built for the Nous Research Plugin Hackathon.

## Post

Introducing  Prompt-Firewall plugin.🪽

A deterministic security layer for AI agents. Detects prompt injections in untrusted tool output and blocks dangerous actions.

Local. No LLM calls. No telemetry.

Built for the @NousResearch Plugin Hackathon. https://github.com/NousResearch/hermes-agent/pull/151850

[Video](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Famplify_video_thumb%2F2120360690103816192%2Fimg%2Fva9thnhs0II_euPC.jpg&mode=thumb&exp=1794700800&sig=407be856e6ff6723)