GPU CLI: Remote GPU Compute for ML Training and Inference

Safely run local `gpu` commands via a guarded wrapper (`runner.sh`) with preflight checks and budget/time caps.

GPU CLI Skill (Stable)

Use this skill to run the local gpu binary from your agent. It only allows invoking the bundled runner.sh (which internally calls gpu) and read-only file access.

What it does

• Runs gpu commands you specify (e.g., runner.sh gpu status --json, runner.sh gpu run python train.py).
• Recommends a preflight: gpu doctor --json then gpu status --json.
• Streams results back to chat; use --json for structured outputs.

Safety & scope

• Allowed tools: Bash(runner.sh*), Read. No network access requested by the skill; gpu handles its own networking.
• Avoid chaining or redirection; provide a single runner.sh gpu … command.
• You pay your provider directly; this may start paid pods.

Quick prompts

• "Run runner.sh gpu status --json and summarize pod state".
• "Run runner.sh gpu doctor --json and summarize failures".
• See templates/prompts.md for more examples.

Security

• Input sanitization: character blocklist (; & | \ ( ) > < $ { }+ newlines) plus subcommand allowlist. Commands are executed via directgpu binary invocation — no shell re-evaluation (bash -c/eval`).
• See SECURITY.md for the full threat model, permission rationale, and version history.

Notes

• For image/video/LLM work, ask the agent to include appropriate flags (e.g., --gpu-type "RTX 4090", -p 8000:8000, or --rebuild).