Give Your Coding Agent a Web-Scraping Department

What it does

The repo ships five Markdown-based skill packs—plus a slash-command pack—that plug into Claude Code, Cursor, Windsurf, Codex, or Gemini CLI. Once loaded, an agent can pick from 130+ curated Apify Actors to scrape major social, search, and commerce sites, fall back to the broader store of 30,000+ Actors, or build and deploy entirely new serverless scrapers in JavaScript, TypeScript, or Python. It also handles the unglamorous but necessary plumbing: generating output schemas from existing code, wrapping legacy scripts as runnable Actors, and calling runs programmatically from JS/TS or Python apps.

The interesting bit

Instead of handing agents a generic API client, the skills encode hard-won domain knowledge about which Actor fits which site, then automate the full lifecycle from prompt to structured data. The apify-ultimate-scraper skill is essentially a smart dispatcher: it maps a vague request like “top Italian restaurants in Brooklyn” to the right Google Maps Actor, chains in review scraping and website crawling, and formats the result—without the agent having to browse the store or read Actor docs.

Key highlights

• apify-ultimate-scraper covers 130+ curated platforms (Instagram, TikTok, Google Maps, Amazon, Yelp, etc.) and searches the full Apify Store for anything missing.
• apify-actor-development scaffolds, debugs, and deploys new Actors using official SDK patterns.
• apify-actorization wraps existing scripts, libraries, or CLI tools into runnable Actors with proper I/O handling.
• apify-generate-output-schema reverse-engineers dataset_schema.json and friends by analyzing Actor source code.
• apify-sdk-integration exposes sync and async clients for JS/TS and Python, with a REST fallback for other languages.

Verdict

Worth a look if you want your agent to move beyond toy examples and actually extract, transform, and ship web data. Skip it if you don’t need scraping or you’d rather orchestrate Actors manually through the Apify Console.