---
format: "aidr-story-markdown/v1"
id: "b4439a7527ca175b45e558714e0c076936f57c7b83f7c3dca55c2d93f6736ea5"
canonical_url: "https://aidr.today/b4439a75?lang=en"
title: "External sandboxes for coding agents"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-11-17T22:52:29.000Z"
category: "Open Source"
topics: ["safety","open-source","openai","devtools"]
source_urls: ["https://simonwillison.net/2026/Nov/17/external-sandboxes-for-coding-agents/","https://openai.com/index/inside-the-codex-sandbox"]
summary: "External sandboxes for coding agents I've been exploring new ways to sandbox code that is executed by agents, with a focus on making it safer to try out local models without worrying that a mistake or prompt injection attack might cause them to take damaging actions on a machine. I worked on this project with Prime Radiant , the AI research lab behind Superpowers and other projects . My post on the Prime Radiant blog introduces the external sandbox pattern, where code executed by the agent runs inside a dedicated sandbox while the agent itself lives outside of that sandbox. To demonstrate this pattern we've released three sandbox server implementations: external-sandbox-v86 runs a Node.js server which then uses v86 to run Alpine Linux inside WebAssembly. external-sandbox-container is a Python web server which runs the sandbox in Alpine Linux inside an Apple Containers instance. external-sandbox-quicksand is another Python web server, this time using Microsoft's Quicksand sandboxing library, which bundles QEMU and Alpine Linux in a Python package! Each of these implementations offers the same localhost API."
---

# External sandboxes for coding agents

> [Open the canonical story](<https://aidr.today/b4439a75?lang=en>)

**Published:** 2026-11-17T22:52:29.000Z
**Category:** Open Source
**Topics:** safety, open\-source, openai, devtools

## Summary

External sandboxes for coding agents I've been exploring new ways to sandbox code that is executed by agents, with a focus on making it safer to try out local models without worrying that a mistake or prompt injection attack might cause them to take damaging actions on a machine\. I worked on this project with Prime Radiant , the AI research lab behind Superpowers and other projects \. My post on the Prime Radiant blog introduces the external sandbox pattern, where code executed by the agent runs inside a dedicated sandbox while the agent itself lives outside of that sandbox\. To demonstrate this pattern we've released three sandbox server implementations: external\-sandbox\-v86 runs a Node\.js server which then uses v86 to run Alpine Linux inside WebAssembly\. external\-sandbox\-container is a Python web server which runs the sandbox in Alpine Linux inside an Apple Containers instance\. external\-sandbox\-quicksand is another Python web server, this time using Microsoft's Quicksand sandboxing library, which bundles QEMU and Alpine Linux in a Python package\! Each of these implementations offers the same localhost API\.

## Sources

- [Story source](<https://simonwillison.net/2026/Nov/17/external-sandboxes-for-coding-agents/>)
- [Story source](<https://openai.com/index/inside-the-codex-sandbox>)

