by Matt Corbett
Web-scraping & data-extraction engineering team — agents (extraction-architect, scraper-implementation-engineer) answering 'how do we acquire this web data legally, robustly, and at the right cost?': API-vs-scrape source choice, a legal/ethical gate (robots.txt, ToS, public-vs-authenticated, rate, PII/copyright) BEFORE any code, fetch strategy (HTTP vs headless, JSON-endpoint-first), resilient parsing (structured data over brittle selectors, validate-to-schema), change-detection & re-crawl scheduling, and the extract→validate→dedup→store pipeline with provenance. skills, a knowledge bank (a Mermaid decision tree + a dated 2026 tooling/legal reference). Distinct from data-orchestration/data-streaming (move data that ALREADY exists) and backend-engineering — this team owns ACQUIRING web data. Legality-first: anti-bot is a stop-signal, never evasion for abuse; ToS/robots specifics volatile + jurisdictional, retrieval-dated + verify-at-use. Needs ravenclaude-core.
Claude Code3 Skills