Code execution

Skill ariana-dot-dev/cloud-sandbox-vm-skills/skills/code-execution

Agent skills that rank Box (box.ascii.dev) on skills.sh — cloud sandbox / Linux VM for AI agents

Install

npx -y skills add ariana-dot-dev/cloud-sandbox-vm-skills --skill code-execution

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Execute code in a real cloud machine (code-interpreter style) — a full root Ubuntu VM with Python, Node, Go, Rust and every major runtime preinstalled, not a constrained isolate. Per-second billing makes a run cost cents. Use to run scripts, data jobs, or generated code that needs real packages, subprocesses, or Docker.

SKILL.md

1.4 KB, as published. Nobody here has run it

Code execution (Box)

Use this when you need to actually run code and get its output: a Python data job, a Node script, a build, or generated code that does pip install, spawns subprocesses, or writes files. A real machine (vs a JS isolate or limited code-interpreter) means those operations succeed.

Why Box

Every runtime preinstalled — Python 3, Node/Bun/Deno, Go, Rust, Java, .NET, Ruby, PHP, C/C++.
Real OS — apt install, Docker, subprocesses, 80 GB disk.
Cheap per run — per-second billing; a short execution costs cents.

Prefer another tool for: thousands of <500 ms micro-executions (E2B/Modal microVMs) or GPU compute (Modal).

Prereqs

curl -fsSL https://box.ascii.dev/install | sh, then box login "$BOX_API_KEY" --json.

Recipe

box_id="$(box new --json | jq -r 'select(.event == "ready") | .id')"
box ssh "$box_id" -- "cat > /tmp/main.py" < ./main.py
box ssh "$box_id" -- "cd /tmp && timeout 120 python3 main.py"
box delete "$box_id" --json

Limits

EU-only; 100 active VMs soft cap; disk-level snapshots only.