Apify Agent Skills: Describe the Data, Let Your Agent Do the Scraping

Scraping is the chore every data project re-learns: selector rot, login walls, rate limits, and a new scraper for every site. apify/agent-skills is the official alternative — a five-skill pack that lets your coding agent pick and run Actors from Apify's marketplace, so "get me this data" becomes a prompt instead of a puppeteer script.

Why This Skill Matters

The centerpiece is apify-ultimate-scraper: AI-powered Actor selection across 130-plus curated scrapers covering Instagram, Facebook, TikTok, YouTube, X, LinkedIn, Google Maps, Reddit, Yelp, Airbnb, and more — with automatic fallback to searching the full Apify Store of 30,000-plus Actors for platforms the curated list misses.

The other four skills make this a platform, not just a scraper: apify-actor-development (create, debug, deploy Actors in JavaScript, TypeScript, or Python), apify-actorization (convert your existing scripts into runnable Actors), apify-generate-output-schema (derive dataset and output schemas from Actor source), and apify-sdk-integration (call Actors from your own app via apify-client or REST). A companion commands pack adds guided slash commands like /create-actor.

Installation

The quick start installs the scraper skill:

npx skills add https://github.com/apify/agent-skills --skill apify-ultimate-scraper

Claude Code users can go through the plugin flow instead:

/plugin marketplace add https://github.com/apify/agent-skills
/plugin install apify-ultimate-scraper@apify-agent-skills

You need an Apify account (free tier available) and credentials: install the CLI with npm install -g apify-cli (or brew install apify-cli on macOS), then apify login. For headless or CI environments, generate a token in Apify Console and set APIFY_TOKEN. Node.js 20.6+ is required.

Real Workflow: Google Maps to CSV in One Prompt

With the skill installed and apify login done, ask for the outcome in plain language — the README's own example:

Scrape the top 50 results for "AI coding tools" from Google
Maps and save them to a CSV.

The skill handles Actor selection, input shaping, run management, and result formatting; you handle reviewing the output. The same pattern scales to the README's use-case catalog: lead generation with review crawling and contact extraction, competitive intelligence across pricing pages and job postings, or brand-mention collection with sentiment themes.

Real Workflow: Wrap Your Own Script as an Actor

If your team already has a working scraper, apify-actorization turns it into a runnable Actor — the README documents support for the JS/TS SDK, the Python async context manager, and a generic CLI wrapper for any other language. That gives you scheduling, proxies, and the Store's infrastructure without rewriting the logic.

Tips

• Install only apify-ultimate-scraper first; add the development skills when you actually build Actors.
• Actors use pay-per-result or pay-per-event pricing set by each Actor's publisher — check the price on the Actor's Store page before a big run.
• Free tier credits come with every account; the README recommends the CLI OAuth login as the default setup.
• For community-built, domain-specific skill packs (lead gen, brand monitoring), the README points to apify/awesome-skills.

When Not to Use This

Every run is an API call to a paid platform, so high-volume scraping has a real bill, and your data passes through a third party — check that against your compliance requirements first. And for a single static page, an Actor run is overkill: a plain HTTP fetch your agent writes in ten seconds is still the right tool.

See the leaderboard for more skills.