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

# Install the agent skill

> Give agents a bounded workflow and a tested REST fallback.

Install the `google-maps-data` skill from the public repository:

```bash theme={null}
npx skills add ricciflow-api/gmapscrawl-agent-skills
```

You can also copy `skills/google-maps-data` into the skill directory supported by your client. Read the [repository README](https://github.com/ricciflow-api/gmapscrawl-agent-skills) for the package layout and tests.

The skill prefers a configured G Maps Crawl MCP server. When MCP is unavailable, its Node.js helper reads `GMSCRAPER_API_KEY` from the environment and accepts a bounded JSON request file:

```json theme={null}
{
  "operation": "places.search",
  "input": {
    "q": "coffee shops in Seattle",
    "page": 1
  },
  "client_request_id": "gmaps-example-search-20260922",
  "wait_seconds": 10
}
```

```bash theme={null}
node skills/google-maps-data/scripts/gmapscraper-request.mjs request.json
```

Use a fresh identifier for new work. The helper performs one API request; it does not silently poll, expand source pages, or export results. Use `jobs.get` and `jobs.results` request documents to continue.

The package includes operation schemas, result semantics, validation tests, and examples. Installing it does not provision an API key or enable unavailable live capabilities.