# labloop

> labloop — cloudcell-labloop. Use this tool when you need to create a secure and reproducible environment for machine learning research, with isolated zones for drivers, trusted, and hostile components. Labloop provides per-user virtual machines and standalone scientific-state MCP servers, ensuring a reliable and consistent setup for agent-driven ML experiments. Ideal for researchers requiring a controlled and isolated space to test and train AI models.

Canonical page: https://skillsregistry.net/skills/cloudcell-labloop  
JSON: https://api.skillsregistry.net/v1/skills/cloudcell-labloop

## Description

Secure, reproducible KVM lab for agent-driven ML research: per-user VMs with driver / trusted / hostile zone isolation, plus five standalone scientific-state MCP servers

## Trust

- **Trust score (0–1):** 0.78
- **Verification tier:** verified
- **Last scanned:** 2026-09-28

## Facts

- **Version:** 1.0.0
- **Skill type:** atomic
- **Execution layer:** container
- **Runtime environment:** vm
- **License:** Apache-2.0
- **Updated:** 2026-09-28

## Source

- **Source listing:** [GitHub](https://github.com/cloudcell/labloop)

## Use it

Resolve this record through the SkillsRegistry MCP server (no auth, read-only):

```
claude mcp add --transport http --scope user skillsregistry https://api.skillsregistry.net/mcp
```

```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "get_skill",
    "arguments": {
      "slug": "cloudcell-labloop"
    }
  }
}
```

REST: `GET https://api.skillsregistry.net/v1/skills/cloudcell-labloop` · pull for local use: `GET https://api.skillsregistry.net/v1/skills/cloudcell-labloop/pull`

---
SkillsRegistry indexes agent skills from public registries and GitHub. Skills we have analysed are scanned with Circle-IR and scored on six dimensions; each listing states its scan coverage. More: https://skillsregistry.net/llms.txt