---
title: "CLRK: Open-Source AI Agent Runtime with gVisor and MitM Guardrails"
source: "Hacker News 热门（Algolia · 中文化）"
source_tier: "t2"
url: "https://github.com/apoxy-dev/clrk"
permalink: "https://feed.astralor.com/items/itm_77dde9b5ceaf36811f1906e1"
published: "2026-07-07T19:45:15.000Z"
score: 70
channel: "ai"
tags: ["AI Agent Security","apoxy-dev","CLRK","gVisor","Kubernetes"]
---

# CLRK: Open-Source AI Agent Runtime with gVisor and MitM Guardrails

## AI commentary

> Although still early-stage, this project has a very pragmatic design—many teams struggle with security when deploying AI agents, and CLRK directly addresses it with gVisor for process isolation and MitM for communication auditing. Worth watching for DevOps and security teams running agents on Kubernetes.

## At a glance

CLRK (Cognitive Loop Runtime for Kubernetes) is an open-source AI Agent runtime that leverages gVisor sandboxing and MitM (Man-in-the-Middle) guardrails to provide secure isolation and trusted execution for agents in Kubernetes environments. Open-sourced on GitHub, it supports cognitive loop architecture, suitable for high-security automated agent deployments.

## Key points

- Lightweight sandbox isolation via gVisor enhances agent runtime security
- Built-in Man-in-the-Middle (MitM) guardrail layer monitors and filters agent communications
- Designed natively for Kubernetes with cognitive loop architecture support

## Full text

Cogntive Loop Runtime for Kubernetes. Contribute to apoxy-dev/clrk development by creating an account on GitHub.

---

Verify against the original before citing; curated by Astral Feed: https://feed.astralor.com/items/itm_77dde9b5ceaf36811f1906e1