AgentSkillsCN

scraper-status

全面概览所有爬虫——涵盖质量评分、提取率以及覆盖缺口。

SKILL.md
--- frontmatter
name: scraper-status
description: Quick overview of all scrapers — quality grades, extraction rates, and coverage gaps.

Scraper Status Overview

Show the health and quality of all supported scrapers by running the validation test suite and summarizing results.

Workflow

Step 1: Run the full validation suite

bash
cd astro-app && npx vitest run test/lib/scraper-validation.test.ts 2>&1

Step 2: Summarize results

Parse the test output and present a table showing for each scraper:

ScraperGradeExtraction RateWeighted RateCritical MissingStatus

Step 3: Identify issues

Highlight:

  • Failing scrapers (grade F or test failures)
  • Degraded scrapers (grade C, or not meeting expected extraction rate)
  • Missing critical fields (title, price_string, price_float)
  • Coverage gaps (portals without test fixtures)

Step 4: If MCP tools are available

When the property-scraper MCP server is running, you can also use:

  • list_supported_portals to get the full list of supported portals
  • get_scraper_mapping to inspect a specific scraper's selectors
  • extract_property to re-run extraction on a fixture for detailed diagnostics

Output format

Present a concise dashboard:

  • Total scrapers, passing, failing
  • Table of all scrapers with grades
  • Action items for any scrapers needing attention