Apify Agent Skills
Apify official skill pack — scrape 130 curated Actors across Instagram, TikTok, LinkedIn, and Google Maps, or build and deploy your own.
Overview
Apify Agent Skills is the official Agent Skills pack from Apify — five skills plus a commands pack that let an AI coding agent scrape the major social, search, maps, and review platforms through Actors, then build, actorize, and deploy Actors of its own.
What is Apify?
Apify is a marketplace and serverless runtime for web scraping and automation. Its unit of work is the Actor — a containerised cloud program with a declared input schema and structured dataset output. The Apify Store hosts more than 30,000 of them, maintained by Apify itself or the community, and each one absorbs the maintenance burden that breaks hand-rolled scrapers: proxy rotation, login walls, layout changes, and rate limits.
The five skills
apify-ultimate-scraper— the universal scraper. It ships a flat Actor index of 130 curated Actors across 14 categories, so the agent picks a known-good Actor instead of guessing: 20 for LinkedIn, 16 for Facebook, 13 each for Instagram and TikTok, 10 for real estate and hospitality, 8 each for cross-platform reviews and web crawling, 7 each for YouTube and Google Maps, 6 SEO tools, 5 each for X/Twitter and contact enrichment, 3 for Google Search and Trends, plus 9 others. Every entry is tieredapify(always preferred) orcommunity(gap filler), and anything not in the index falls back to a search across the full Apify Store.apify-actor-development— create, debug, and deploy Actors from scratch in JavaScript, TypeScript, or Python, with bundled references foractor.json, input schema, output schema, dataset schema, key-value store schema, logging, and standby mode.apify-actorization— wrap existing code as an Actor, covering the JS/TS SDK, the Python async context manager, and a generic CLI wrapper for any other language.apify-generate-output-schema— derivedataset_schema.json,output_schema.json, andkey_value_store_schema.jsonby reading an Actor's own source.apify-sdk-integration— call Actors from an existing JavaScript/TypeScript or Python application throughapify-client, covering sync and async execution, dataset and key-value store retrieval, error handling, and the REST fallback for other languages.
Alongside them, the apify-actor-commands pack adds slash commands such as /create-actor for guided scaffolding.
Fourteen workflow guides
The scraper skill loads a matching playbook when a task spans multiple steps: lead generation, contact enrichment, company research, competitive intel, brand monitoring, review analysis, social media analytics, influencer vetting, trend research, content and SEO, ecommerce price monitoring, real estate and hospitality, job market and recruitment, and knowledge base and RAG ingestion.
Cost discipline built in
Apify Actors bill per result or per event, so the scraper ships a gotchas.md covering pricing models, a cost estimation protocol to run before a large job, common pitfalls, error recovery, platform-specific rate limits, and a straight argument for when Actors beat raw HTTP scraping. It also fixes three CLI habits that otherwise break automation: always pass --json for output stable across CLI versions, always redirect stderr with 2>/dev/null because progress messages corrupt JSON parsing, and pass --user-agent apify-agent-skills/apify-ultimate-scraper so runs are attributed to the skill.
Who it is for
Growth, research, and data teams who want an agent that can actually pull structured data off Instagram, LinkedIn, Google Maps, Amazon, or Booking.com — and developers packaging their own scrapers as deployable, billable Actors.
What you can build
Ranked lead lists from Google Maps and enriched with owner emails, competitive sweeps across pricing pages and G2 reviews, brand sentiment reports from 30 days of social mentions, influencer shortlists ranked by engagement-to-follower ratio, price-monitoring pipelines across Amazon and Walmart, AI-search visibility checks, and RAG knowledge bases ingested from crawled sites.
Why it matters
Scraping breaks constantly, and an agent writing its own fetch loop will hit a login wall, a rendered-JS shell, or a rate limit and quietly return partial data. Routing through maintained Actors moves that problem to someone whose job it is to fix it — and pinning the agent to a curated index of 130 known-good Actors keeps it from picking a stale community port of the one it needed.
What's Included
- apify-ultimate-scraper — universal scraping skill with Actor selection, input shaping, run management, and result formatting
- An Actor index of 130 curated Actors across 14 categories, each tiered as Apify-maintained or community-maintained
- Automatic fallback to a search across the full Apify Store of 30,000+ Actors when a platform is not in the index
- 14 workflow guides: lead generation, contact enrichment, company research, competitive intel, brand monitoring, review analysis, social analytics, influencer vetting, trend research, content and SEO, ecommerce price monitoring, real estate, recruitment, and RAG ingestion
- gotchas.md — pricing models, a cost estimation protocol, common pitfalls, error recovery, and platform rate limits
- apify-actor-development — build, debug, and deploy Actors in JavaScript, TypeScript, or Python
- Actor development references for actor.json, input schema, output schema, dataset schema, key-value store schema, logging, and standby mode
- apify-actorization — wrap existing code as an Actor via the JS/TS SDK, the Python async context manager, or a generic CLI wrapper
- apify-generate-output-schema — derive dataset, output, and key-value store schemas from an Actor source
- apify-sdk-integration — call Actors from your app with apify-client, covering sync and async runs, storage retrieval, and the REST fallback
- apify-actor-commands — slash commands including /create-actor for guided scaffolding
- agents/AGENTS.md — an auto-generated one-page index of every skill for Markdown-reading agents
- Plugin manifests for Claude Code, Cursor, and Windsurf, plus gemini-extension.json for Gemini CLI auto-discovery
Installation
Option 1 — skills CLI
Install the scraper on its own:
npx skills add https://github.com/apify/agent-skills --skill apify-ultimate-scraperOption 2 — Claude Code plugin
The repository is its own plugin marketplace, and each skill installs separately:
/plugin marketplace add https://github.com/apify/agent-skills
/plugin install apify-ultimate-scraper@apify-agent-skills
/plugin install apify-actor-development@apify-agent-skills
/plugin install apify-actorization@apify-agent-skills
/plugin install apify-generate-output-schema@apify-agent-skills
/plugin install apify-sdk-integration@apify-agent-skillsCursor and Windsurf read the same plugin format.
Option 3 — Codex CLI, Gemini CLI, or any Markdown-reading agent
The repo ships agents/AGENTS.md as a one-page auto-generated index of every skill, and gemini-extension.json so Gemini CLI auto-discovers it. Clone and point your agent at it:
git clone https://github.com/apify/agent-skillsOtherwise reference skills/<skill-name>/SKILL.md directly.
Set up the Apify CLI
npm install -g apify-cli # or: brew install apify-cli
apify login # OAuth in your browser — the recommended pathFor headless or CI environments where no browser is available, generate a token in Apify Console under Integrations and set it instead:
export APIFY_TOKEN="your-token-here"The CLI picks up APIFY_TOKEN from the environment or a .env file automatically. The Ultimate Scraper skill needs Apify CLI v1.5.0 or newer and Node.js 20.6+.
Use it
Describe the outcome and let the skill choose the Actor:
Find Italian restaurants in Brooklyn rated under 4 stars on Google Maps, scrape their reviews, crawl their sites for owner emails, and export a ranked CSV.
The skill reads its Actor index to pick the right Actor, loads the matching workflow guide for multi-step jobs, shapes the input, manages the run, and formats the results.
Optional — hosted MCP
mcp.apify.com generates a one-line MCP URL exposing the Actors you choose as tools, for any MCP-compatible client.
Requirements
- An AI agent that reads Agent Skills or the Claude Code plugin format — Claude Code, Cursor, Windsurf, Codex CLI, or Gemini CLI
- An Apify account (free tier available) at apify.com
- Apify CLI v1.5.0 or newer, installed with npm install -g apify-cli or brew install apify-cli
- Authentication via apify login (browser OAuth), or APIFY_TOKEN in your environment or .env for headless and CI use
- Node.js 20.6 or newer, required by the Apify CLI and the Ultimate Scraper skill
- Apify credits — Actors bill per result or per event, with free tier credits included on every account
Changelog
Initial marketplace listing of the official Apify agent skills pack (5 skills plus a commands pack, 130 curated Actors), matching plugin manifest version 2.0.0.
FAQs
Related Skills
Firecrawl Skills
Firecrawl official skill catalog — 33 skills for live web search, scrape, crawl, and interact, plus SDK integration and workflow recipes.
XLSX
Create, edit, and fix Excel spreadsheets — .xlsx, .xlsm, .csv, .tsv — with live formulas, formatting, and charts, zero formula errors.