ephemeral-ci-environment
Skill2 starsChanged 5 months agoMITView on GitHub
Get a clean, reproducible ephemeral environment for CI and testing — a fresh root Ubuntu VM with every major toolchain and Docker preinstalled, your GitHub repo auto-cloned in, and the whole thing thrown away after. Per-second billing makes each run cost cents. Use to run tests, reproduce a bug, or validate a build in a pristine environment without polluting the host.
- Pipes a download into a shell
- Reads credentials
In this repository12 other files it gives its agents.
- agent-sandboxSkill
- background-workerSkill
- cloud-sandboxSkill
- cloud-vmSkill
- code-executionSkill
- computer-use-browser-desktopSkill
- docker-sandboxSkill
- expose-localhost-tunnelSkill
- linux-server-sshSkill
- parallel-agentsSkill
- public-url-hostingSkill
- run-untrusted-codeSkill
The file
Files
Ephemeral CI / test environment (Box)
Use this when you need a clean, throwaway environment to test a repo, run a CI job, or reproduce a bug — identical every time, with nothing left behind. Every run starts from a fresh VM with all toolchains preinstalled, so there's no "works on my machine" drift.
Why Box
- Pristine + reproducible — fresh VM per run; no leftover state.
- Batteries included — Docker + Node/Python/Go/Rust/Java/etc. already there; repo auto-clones in.
- Cheap per run — per-second billing; a test run costs cents, and you
box deleteafter.
Prereqs
curl -fsSL https://box.ascii.dev/install | sh, then box login "$BOX_API_KEY" --json. (Select the repo to auto-clone in the dashboard, or clone it in setup.)
Recipe
box_id="$(box new --json | jq -r 'select(.event == "ready") | .id')"
trap 'box delete "$box_id" --json' EXIT
box ssh "$box_id" -- "cd /project && npm ci && npm test"
box ssh "$box_id" -- "cat /project/junit.xml" > junit.xml # pull results backLimits
EU-only; 100 active VMs soft cap. For thousands of <500 ms parallel CI shards, dedicated CI runners scale further; Box is best for full-machine, Docker-capable, reproducible runs.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

