I built an open MVP for deploying AI-built software without giving agents unrestricted server access

Описание и применение

Guardian Cloud – открытый MVP, позволяющий безопасно развертывать приложения, созданные ИИ, без предоставления агенту неограниченного доступа к серверу. Поддерживает публичные репозитории Node.js, предоставляет план развертывания, требует явного одобрения и обеспечивает автоматическое доменное имя и HTTPS.

Задача

Deploying AI-generated applications to production servers is difficult: either a human does all steps manually, which doesn't scale, or the AI agent is given unrestricted shell access, risking mistakes.

Результат

Guardian Cloud successfully deployed two external Node.js repositories: one reached a public HTTPS URL with GET / returning 200, the other failed honestly at the build step, demonstrating fail‑closed behavior.

Рабочий процесс

• Paste public GitHub repo URL
• Guardian analyzes the app and detects Node.js (user provides port if needed)
• Guardian presents deployment plan
• User explicitly approves deployment
• Guardian performs controlled infrastructure actions, monitors deployment, provisions domain and HTTPS
• Application is deployed, validated, and exposed via public URL

Доказательства

Here is what the last validation round actually showed: an external Node.js sample repository was deployed and later served GET / = 200 publicly; another external Node.js repository (openshift/nodejs-ex) reached the deployment stage and failed honestly at the build step; the final end‑to‑end LIVE test used a small public Node.js test repository (AndersonVitaease/guardian-cloud-e2e) and reached a public HTTPS URL with GET / = 200.