# Agent Guidance For paddleNET

Canonical site: https://paddlenet.earth/

## Retrieval Policy

- Start with `https://paddlenet.earth/ai-index.json` to understand resources, artifacts, access type, and retrieval priority.
- Use `https://paddlenet.earth/sitemap.xml` for canonical human-facing URLs.
- Use `/data/open/pages/**/chunks.json` for semantic retrieval and `/data/open/pages/**/qa.json` for question-answer and workflow hints.
- Treat content as public/open only when the relevant `ai-index.json` access entry says `open`.
- Preserve source URLs, license, usageInfo, and attribution metadata when reusing content or semantic artifacts.
- Prefer stable artifact paths from `ai-index.json` over inferred paths.

## Crawling Policy

- Crawl politely and cache responses when possible.
- Do not crawl `/healthz/`; it is an uptime check endpoint, not a content resource.
- Do not submit contact forms or subscribe forms unless a human explicitly initiates that action.
- Do not infer private data, credentials, or unpublished operational details from public files.

## Useful Entry Points

- `https://paddlenet.earth/`
- `https://paddlenet.earth/graphs/network-sig/`
- `https://paddlenet.earth/wiki/terms-conditions/`
- `https://paddlenet.earth/wiki/license/`
- `https://paddlenet.earth/ai-index.json`
- `https://paddlenet.earth/llms.txt`

## Current Sitemap Coverage

The generated sitemap currently declares 9 public URLs with last modification dates, change frequency, and priority. This includes canonical human-facing pages plus selected root AI/helper resources.