deploy-modal
Use when deploying Modal workers, checking Modal logs, managing Modal secrets, or configuring GPU endpoints. Triggers on modal deploy, Modal worker, Modal app, Modal logs, Modal secret, modal run, Modal endpoint, keep_warm, cold start, Modal proxy token, wk- ws- token, 8-endpoint limit, Modal GPU, memory sizing, deploy_all.py, Modal Python worker.
Source facts
- Repository
- george11642/george-plugins
- Last source activity
- March 11, 2026 at 22:42
- Detected SKILL.md language
- English
- Stars
- 2
- Forks
- 0
Install options
The review-first prompt is selected by default. You can switch to a direct command or download a local copy.
Review the source files
Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.
Showing SKILL.md
SKILL.md
Source instructions · Read-only preview- name
- deploy-modal
- description
- Use when deploying Modal workers, checking Modal logs, managing Modal secrets, or configuring GPU endpoints. Triggers on modal deploy, Modal worker, Modal app, Modal logs, Modal secret, modal run, Modal endpoint, keep_warm, cold start, Modal proxy token, wk- ws- token, 8-endpoint limit, Modal GPU, memory sizing, deploy_all.py, Modal Python worker.
# Deploy Modal
Check project CLAUDE.md for worker paths and deploy entry points.
## Commands
| Task | Command |
|------|---------|
| Deploy app | `modal deploy path/to/app.py` |
| List apps | `modal app list` |
| View logs | `modal app logs app-name` |
| Set secret | `modal secret create name KEY=value` |
| Run ad-hoc | `modal run path/to/script.py` |
## Gotchas
- **8-endpoint limit** — multi-route workers consolidate endpoints
- **Proxy auth tokens** (`wk-*`/`ws-*`) differ from CLI tokens (`ak-*`/`as-*`) — Convex env vars use proxy tokens
- **Memory sizing**: right-size per worker to reduce costs
- **Cold starts**: use `keep_warm=1` for latency-sensitive endpoints
## Worker Pattern
```python
app = modal.App("worker-name")
image = modal.Image.debian_slim().pip_install(...)
@app.function(image=image, memory=512)
@modal.web_endpoint(method="POST")
def endpoint(request: dict):
...
```
View on GitHub