runpod

# runpod

RunPod GPU cloud management plugin for Claude Code.

## Installation

```bash
claude plugins add /path/to/runpod
```

## Skills

### prepare-model-upload

Download HuggingFace models on Colab and upload to RunPod Network Volume via S3-compatible API.

Pre-deploy models before GPU instance startup to save billing time.

#### Prerequisites

Set these in Colab Secrets:
- `HF_TOKEN`: HuggingFace access token
- `RUNPOD_STORAGE_ACCESS_KEY_ID`: RunPod Storage Access Key ID
- `RUNPOD_STORAGE_SECRET_ACCESS_KEY`: RunPod Storage Secret Access Key

#### Usage

1. Open the Colab notebook
2. Configure Colab Secrets
3. Enter model names and RunPod Storage settings, then run

### create-pod

Create RunPod GPU pod instances from a `runpod.toml` config file.

#### Prerequisites

- `runpodctl` CLI installed and configured (`~/.runpod/config.toml`)

#### Usage

1. Copy template: `cp <plugin>/skills/create-pod/templates/runpod.toml ./runpod.toml`
2. Edit `runpod.toml` with your settings
3. Create pod: `uv run --script create_pod.py`
4. Create and SSH: `uv run --script create_pod.py --ssh`

#### Features

- Config-driven pod creation (`runpod.toml`)
- RunPod Secret support (`{{ RUNPOD_SECRET_XXX }}`)
- SSH auto-connect with init script and tmux session
- CLI overrides for GPU, datacenter, pod name
- `--dry-run` to preview commands

### volume-storage

Manage files on RunPod Network Volume via S3-compatible API.

#### Prerequisites

- `aws` CLI installed with `[profile runpod]` in `~/.aws/config`
- `runpod.toml` with `datacenter_id` and `network_volume_id`

#### Usage

```bash
volume_storage.sh ls s3://VOLUME_ID/
volume_storage.sh cp s3://VOLUME_ID/file.json ./file.json
volume_storage.sh sync ./output/ s3://VOLUME_ID/output/
```

### stocks

Check RunPod GPU availability and suggest GPUs/datacenters matching requirements.

#### Prerequisites

- `~/.runpod/config.toml` with API key, or `RUNPOD_API_KEY` environment variable

#### Filter Options

- `--min-memory`: Minimum VRAM (GB)
- `--gpu`: Filter by GPU name
- `--storage`: Network Volume supported only
- `--stock`: Stock level (high/medium/low)
- `--gen`: GPU generation (blackwell/hopper/ada/ampere/volta/amd)