smolvm
Sub-200ms microVMs for sandboxing AI coding agents safely
The Panel's Take
smolvm is a lightweight microVM runtime built in Rust on top of libkrun, designed specifically for sandboxing AI coding agents and untrusted code execution. VMs cold-start in under 200ms and ship as portable `.smolmachine` files — think Docker images but hardware-isolated. It supports macOS (Apple Silicon and Intel) and Linux, with opt-in networking so that untrusted code can't exfiltrate credentials or phone home by default. The project includes an explicit AGENTS.md to help coding agents understand how to use it, and was built with autonomous code execution in mind. When an AI agent needs to run user-submitted code or iterate on its own suggestions, smolvm gives it a proper hardware sandbox rather than a leaky container. Version v0.5.18 landed April 17, 2026. With AI coding agents increasingly running arbitrary code in tight loops, the security story around containerization has become critical. smolvm fills a real gap: fast enough to not break agentic workflows, isolated enough to actually protect the host machine and credentials. It surfaced on Hacker News with 259 points and strong technical discussion, suggesting genuine resonance with the developer community building agentic tools.
Share this verdict
smolvm verdict: SHIP 🚀 3 ships · 1 skip from the expert panel Full review: shiporskip.io/tool/smolvm-microvm-ai-agent-sandboxing-rust-libkrun-sub-200ms-2026
Weekly AI Tool Verdicts
Get the next verdict in your inbox
7 critics review a new AI tool every day. Weekly digest — free.
Compare smolvm with Others
Embed this verdict
Tool makers can add a live ShipOrSkip badge to their site. Badge loads track impressions; clicks route back to this review.
<a href="https://shiporskip.io/api/badge-click/smolvm-microvm-ai-agent-sandboxing-rust-libkrun-sub-200ms-2026" target="_blank" rel="noopener"><img src="https://shiporskip.io/api/badge/smolvm-microvm-ai-agent-sandboxing-rust-libkrun-sub-200ms-2026" alt="smolvm Ship verdict on ShipOrSkip" width="360" height="90" /></a>[](https://shiporskip.io/api/badge-click/smolvm-microvm-ai-agent-sandboxing-rust-libkrun-sub-200ms-2026)<iframe src="https://shiporskip.io/embed/smolvm-microvm-ai-agent-sandboxing-rust-libkrun-sub-200ms-2026" title="smolvm ShipOrSkip verdict" width="360" height="260" style="border:0;border-radius:16px;max-width:100%;" loading="lazy"></iframe>The reviews
“This is the missing layer for anyone running AI agents that execute code. Docker containers have always been too porous for untrusted execution, and smolvm's sub-200ms coldstart means you can spin a fresh VM per agent turn without killing your latency budget. The AGENTS.md is a thoughtful touch — shows the authors actually understand the workflow.”
“At v0.5.18 this is still early software and the docs are sparse. libkrun has its own surface area of bugs, and running microVMs at agent-loop speed on macOS introduces a whole class of Apple Hypervisor Framework edge cases. I'd wait for v1.0 and a production case study before betting real workloads on this.”
“Every autonomous agent that executes code needs a proper sandbox — not a polite request for the agent to be careful. smolvm represents the infrastructure layer that makes truly autonomous code execution safe enough to deploy at scale. This kind of primitive is foundational for the agentic software era.”
“For anyone building AI tools that touch code, smolvm means you can let your AI actually run things without fear. That unlocks a whole category of 'show me the output' UX patterns that weren't safe before. Less time explaining sandboxing to users, more time shipping features.”