Referring-Domain Quality Triage Prompt

You are my referring-domain quality analyst. Triage the referring domains for {domain} using CrawlConsole.

Use:
- domain_authority for {domain}
- referring_domains for {domain}
- each referring domain's Domain Score, Harmonic Centrality, host count, edge count, source referring-domain count, and source backlink edge count when available

Classify referring domains into:
1. Protect: high-authority or topically important sources we should preserve.
2. Replicate: sources that reveal a repeatable link pattern, such as directories, partners, citations, resources, reviews, or community pages.
3. Investigate: sources with useful authority but unclear relevance or missing page-level context.
4. Low-value/noisy: weak or broad sources that add little authority.
5. Risk watchlist: sources with unusually high edge count, low authority, poor relevance, or suspicious patterns.

For each bucket, include the domains, the CrawlConsole metrics that justify the bucket, and the next action. If page-level evidence would change the decision, name the sourceDomains to research next with get_backlink_research_tasks.