# Vibeinfra Lab Runner

> Orchestrate disposable sandbox provisioning, execute diagnostic workstation probes, and trigger automated 7-step evaluation grading for VibeInfra labs.

- Skill: `vibeinfraid/vibeinfra-lab-runner` (Agent Skill)
- Install (CLI): `npx skillmds@latest add vibeinfraid/vibeinfra-lab-runner`
- Raw SKILL.md: https://api.skillmd.com/api/skills/vibeinfraid/vibeinfra-lab-runner/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: vibeInfraid (https://skillmd.com/u/vibeinfraid)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/vibeinfraid/vibeinfra-lab-runner

---


# VibeInfra Lab Runner Skill

Use this skill to automate sandbox lifecycle management and execute grading evaluators for VibeInfra interactive labs.

## Workflow

1. **Course Discovery**:
   - Query `GET https://api.vibeinfra.id/api/v1/courses` or read `https://vibeinfra.id/openapi.json`.
   - Select target course ID (e.g., `incident-21-flight-simulator`).

2. **Sandbox Provisioning**:
   - Call `POST https://api.vibeinfra.id/api/v1/sandbox/start` with payload `{"course_id": "..."}`.
   - Wait for container initialization (Mode 1 single-container ~5s, Mode 2 multi-container cluster ~30s).
   - Capture `session_id` and WebSocket terminal connection credentials.

3. **Live Terminal Diagnostics**:
   - Connect to interactive PTY terminal via WebSocket `wss://api.vibeinfra.id/api/v1/sandbox/ws?session_id=<id>`.
   - Run non-destructive inspection commands (`curl -i`, `docker ps`, `kubectl get pods`, `dmesg | tail`).

4. **Grading & Verification**:
   - Call `POST https://api.vibeinfra.id/api/v1/sandbox/eval` with payload `{"session_id": "..."}`.
   - Inspect JSON checkpoint results and verify score.