Data Scraper Agent

This is a complete starter kit for building scheduled scrapers that collect public data and let an LLM score or summarize the results. It's architected around GitHub Actions for free hosting, Gemini Flash for free inference with automatic model fallback when you hit rate limits, and storage adapters for Notion, Sheets, or Supabase. The batch processing approach keeps you inside free tier quotas, and there's a feedback loop so the agent learns from your decisions over time. Good for job monitoring, price tracking, or any repetitive data collection task where you want AI enrichment without paying for infrastructure. The codebase includes concrete source connectors, a learning memory system, and backfill scripts for existing data.

Install to Claude Code

npx -y skills add affaan-m/ecc --skill data-scraper-agent --agent claude-code