hybrid-scrape-stack

Use this skill when a source is complex enough that "just use Playwright" would be too expensive, too brittle, or too slow.

Read First

• AGENTS.md
• data/source_registry.json
• agent-skills/browser-scrape-ops/SKILL.md
• agent-skills/runtime-compliance-evaluator/SKILL.md
• scripts/live_scraper.py

Principle

Treat scraping as a ladder:

1. Plain HTTP and HTML or JSON parsing
2. Browser-assisted network tracing
3. Replayed XHR, GraphQL, or JSON API requests
4. Managed browser or extraction platforms only when the lower tiers fail or cost more to maintain