> ## Documentation Index
> Fetch the complete documentation index at: https://docs.compute.cx/llms.txt
> Use this file to discover all available pages before exploring further.

# Agent Skill

> Give Cursor, Claude Code, or Codex a spend-aware Compute workflow.

Give a coding agent the Compute Agent Skill so it installs the CLI, dry-runs a real example, and only then submits a GPU job.

## Raw skill (for agents)

Agents should fetch this file:

[https://compute.cx/SKILL.md](https://compute.cx/SKILL.md)

```bash theme={null}
curl -fsSL https://compute.cx/SKILL.md
```

<Warning>
  GPU runs spend prepaid credit. `--yes` skips the interactive price confirmation and still bills. `--dry-run` prints the upload plan and creates **no** machine.
</Warning>

## What to tell the agent

```text theme={null}
Follow https://compute.cx/SKILL.md. Dry-run first. Do not pass --yes until I confirm spend.
```

That is enough for Cursor, Claude Code, Codex, and similar tools that can fetch a URL.

## What the skill does

1. Installs with `curl -fsSL https://compute.cx/install.sh | sh` and `compute setup`.
2. Writes the tested `simple_mlp.py` example from [First run](/get-started/first-run).
3. Runs `compute run simple_mlp.py::train --gpu H100-SXM --dry-run` (no machine).
4. On your confirmation, submits with `--wait --yes` and checks teardown via `compute machines`.

Do not paste API keys, `config.toml`, or Checkout session ids into the agent chat.

## Next

[First run](/get-started/first-run) is the same example, written for a human. [Run](/cli/run) is the flag reference.