9 items tagged “sandboxing”
2024
How Anthropic built Artifacts. Gergely Orosz interviews five members of Anthropic about how they built Artifacts on top of Claude with a small team in just three months.
The initial prototype used Streamlit, and the biggest challenge was building a robust sandbox to run the LLM-generated code in:
We use iFrame sandboxes with full-site process isolation. This approach has gotten robust over the years. This protects users' main Claude.ai browsing session from malicious artifacts. We also use strict Content Security Policies (CSPs) to enforce limited and controlled network access.
Artifacts were launched in general availability yesterday - previously you had to turn them on as a preview feature. Alex Albert has a 14 minute demo video up on Twitter showing the different forms of content they can create, including interactive HTML apps, Markdown, HTML, SVG, Mermaid diagrams and React Components.
The first four Val Town runtimes
(via)
Val Town solves one of my favourite technical problems: how to run untrusted code in a safe sandbox. They're on their fourth iteration of this now, currently using a Node.js application that launches Deno sub-processes using the node-deno-vm npm package and runs code in those, taking advantage of the Deno sandboxing mechanism and terminating processes that take too long in order to protect against while(true) style attacks.
2023
In general my approach to running arbitrary untrusted code is 20% sandboxing and 80% making sure that it’s an extremely low value attack target so it’s not worth trying to break in.
Programs are terminated after 1 second of runtime, they run in a container with no network access, and the machine they’re running on has no sensitive data on it and a very small CPU.
Cyber (via) “Cyber is a new language for fast, efficient, and concurrent scripting.” Lots of interesting ideas in here, but the one that really caught my eye is that its designed to be easily embedded into other languages and “will allow the host to insert gas mileage checks in user scripts. This allows the host to control how long a script can run”—my dream feature for implementing a safe, sandboxed extension mechanism! Cyber is implemented using Zig and LLVM.
Python Sandbox in Web Assembly (via) Jim Kring responded to my questions on Mastodon about running Python in a WASM sandbox by building this repo, which demonstrates using wasmer-python to run a build of Python 3.6 compiled to WebAssembly, complete with protected access to a sandbox directory.
2020
Sandboxing and Workload Isolation (via) Fly.io run other people’s code in containers, so workload isolation is a Big Deal for them. This blog post goes deep into the history of isolation and the various different approaches you can take, and fills me with confidence that the team at Fly.io know their stuff. I got to the bottom and found it had been written by Thomas Ptacek, which didn’t surprise me in the slightest.
2010
premasagar’s sandie. “Sandie is a simple method for loading external JavaScript files into a page without affecting the global scope, to avoid collisions between conflicting scripts”—works by loading the script in an invisible iframe (hence a new global scope) and then passing a reference to a callback function in the parent page.
2008
Secure mashups with dojox.secure (via) dojox.secure is brilliant and terrifying at the same time. It provides a full featured API for running untrusted JavaScript in a sandbox, by parsing and validating that code against a variant of Douglas Crockford’s ADsafe JavaScript subset. It could be fantastically useful, but it’s difficult to judge how secure this approach really is.
YUI 3.0 Preview Release 1. YUI sandboxing is a really good idea, which cleverly addresses both the need to run multiple versions of the library at once and the complaints about how verbose traditional YUI code can get.