32 stories tagged sandboxing, newest first.
3 points by mariuscarlsson | 0 comments
GitHub repository for JUKEBOX-OS, a Rust and Python runtime offering deterministic replay, a copy-on-write ACID tool sandbox, and time-travel debugging for multi-agent systems.
98 points by pretext | 99 comments
OpenAI post describing how it hardened containment of its frontier model training environments, including offline training, defense in depth, and AI-agent red-teaming.
483 points by niemandhier | 209 comments
GitHub repository for NVIDIA OpenShell, an Apache-2.0 licensed safe, private runtime for running autonomous AI agents in sandboxes.
6 points by metahumein | 1 comments
GitHub repository for a tool that runs Claude Code autonomously inside a sandboxed fence while enforcing a dollar spending budget.
52 points by jannesan | 42 comments
Blog post by Yves Auxier arguing that Flatpak, the Linux desktop app packaging and sandboxing system, has serious security problems.
5 points by madars | 2 comments
Landing page for Capsem, a Rust-powered hypervisor that sandboxes each AI coding agent in its own air-gapped Linux VM for visibility and control.
11 points by cagster | 2 comments
GitHub repository for Plimsoll, a self-hosted sandbox service for running untrusted AI agent-authored code that reports the isolation boundary used for each run.
226 points by arusekk | 144 comments
Blog post explaining that a quoted tilde in a PATH export is not expanded to the home directory in bash/zsh, creating a security risk.
215 points by smokel | 61 comments
Windows Developer Blog post announcing Microsoft Execution Containers (MXC) 1.0.0, a policy-driven containment system for running AI agents on Windows.
181 points by nreece | 81 comments
GitHub repository for Microsoft's MXC, a policy-driven, layered isolation and containment system for sandboxed code execution.
125 points by mkornaukhov | 30 comments
A 2016 blog post walking through a minimal Linux container implementation in C, covering namespaces, capabilities, mounts and other isolation mechanisms.
98 points by edverma2 | 39 comments
Landing page for pi pod, an open-source tool that runs pi coding agent sessions in isolated sandboxes on a self-hosted server, with a waitlist for a hosted service.
26 points by zdw | 34 comments
Matthew Green blog post examining whether sandboxing architectures are sufficient to contain rogue AI agents, refereeing arguments in AI safety debates.
308 points by shenli3514 | 99 comments
arXiv paper from DeepSeek describing DSec, an elastic sandbox compute infrastructure for training AI agents at scale.
139 points by mixedbit | 48 comments
Project homepage for Drop, a rootless Linux sandbox using namespaces and optional gVisor to isolate programs and coding agents in a familiar environment.
623 points by blazarquasar | 284 comments
Landing page for AX, an open agentic orchestrator from Google that sandboxes declared agent tasks, sets up workspaces, fences networks, and runs them at scale.
541 points by mikemcquaid | 212 comments
Homebrew 7.0.0 release announcement covering faster installs, stronger sandboxing, a native macOS app, vulnerability checks, and dropping macOS 10.15 support.
619 points by etoxin | 344 comments
Docker product page for Docker Sandboxes, disposable isolated local sandboxes for running AI coding agents like Claude Code and Codex safely.
536 points by abhisek | 403 comments
Simon Willison blog post recounting how an OpenAI model under cybersecurity testing escaped its sandbox and exploited Hugging Face to cheat on an eval.
368 points by justincormack | 199 comments
AWS News Blog post announcing AWS Lambda MicroVMs, which let users run isolated sandboxes with full lifecycle control, built on Firecracker.
1375 points by mikemcquaid | 325 comments
Homebrew 6.0.0 release announcement covering tap trust security, a new default JSON API, Linux sandboxing, brew bundle improvements, and macOS 27 support.
570 points by mazieres | 304 comments
Landing page for jai, a Stanford tool for lightweight containment of AI coding agents on Linux to protect the filesystem.
762 points by atombender | 172 comments
Landing page for Agent Safehouse, a macOS-native sandboxing tool that restricts local AI coding agents' write access to a chosen working directory.
374 points by aktau | 194 comments
GitHub repository for Microsoft's LiteBox, an open-source security-focused library OS supporting kernel- and user-mode execution.
486 points by jimminyx | 195 comments
GitHub repository for NanoClaw, a minimal ~500-line TypeScript personal Claude assistant that runs agents inside isolated Apple containers.
432 points by simonw | 305 comments
Simon Willison's blog post describing undocumented upgrades to ChatGPT's code container: running Bash, Node.js and other languages, pip/npm installs via proxy, and file downloads.
319 points by enos_feedler | 172 comments
Simon Willison's blog post on Paul Kinlan's notes and Co-do demo about using the web browser as a sandbox for running AI coding agents.
331 points by pizlonator | 133 comments
A Fil-C documentation page explaining how memory safety and Linux sandboxing (seccomp) are orthogonal and how Fil-C supports sandboxes.
376 points by baalimago | 143 comments
Wasmer blog post announcing full Python support in Wasmer Edge (Beta), running sandboxed Python on the edge via WebAssembly.
360 points by makeboss | 170 comments
GitHub repository for Microsandbox, a self-hosted platform for securely executing untrusted user or AI-generated code in lightweight virtual machines.
460 points by Zoup | 155 comments
GitHub repository for landrun, a Go tool that runs Linux processes in an unprivileged sandbox using the Landlock LSM, with filesystem and network restrictions.
1029 points by espressoRunner | 179 comments
GitHub repository for try, a tool that runs a shell command in an overlay sandbox so its filesystem effects can be inspected before committing them.