tpu-skill-claude

xbill9/tpu-skill-claude

★ 1 stars Python AI/LLM Updated today

Claude Code skill + MCP server for managing Google Cloud TPU capacity and Gemma 4 vLLM serving on TPU VMs

View on GitHub → 🔍 Audit Wallet Slippage →

Quick Install

Copy the config for your editor. Some servers may need additional setup — check the README.

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "tpu-skill-claude": {
      "command": "uvx",
      "args": [
        "tpu-skill-claude"
      ]
    }
  }
}

Or install with pip: pip install tpu-skill-claude

README Excerpt

This repository packages the **`tpu-management` Claude Code skill** and the **`tpu-devops` MCP server**: an AI DevOps/SRE agent for operating Google Cloud TPU capacity and **Gemma 4** vLLM serving on TPU VMs. It finds and provisions TPU capacity (flex-start VMs, queued resources), starts and debugs vLLM, verifies model health, runs benchmarks, analyzes logs with the self-hosted Gemma 4 model, and tears everything down safely.

Tools (1)

Makefile

tpu-skill-claude

Quick Install

README Excerpt

Tools (1)

Related AI/LLM Servers