Gpu Server Management

Use when you need a GPU server to train, infer, or run any compute task — to connect to one, claim and release tasks with countdown timers, track server physical remaining lifespan, execute the 6-step Server Download Strategy (Workers RAG -> anysearch/Cloudflare browser -> Direct vs Multi-Proxy speed benchmark & dynamic race -> Multi-Proxy chunk-aggregated downloading for >500MB -> local machine fallback upload -> register_dataset), query the unified Troubleshooting RAG knowledge base upon encountering ANY issue or error (query_troubleshooting), record and query server-specific pitfalls and caveats (record_pitfall / remove_pitfall / pitfalls memory in get_servers), import/manage Clash & V2Ray proxy subscriptions via import_proxy_subscription, trigger intelligent dual-mode backups (outputs-only vs. full evacuation) with cloud RAG vector indexing, query cached data via query_backup_index / everything-mcp and Dataset Affinity, borrow disk from another machine, or route around a slow or blocked network. Also use

ixijxjgxidj-cmd Updated

File contents

ixijxjgxidj-cmd/GPU-Server-Management-MCP-Skill-/tree/main/.dsh/skills/gpu-server-management commit 60d4212274

Frequently asked questions

npx skillmds@latest add ixijxjgxidj-cmd/gpu-server-management