Skip to content
Get packs

Guide · Agent skill

How to use the Apify Actor Builder agent skill

Updated · written with AI from the project's own documentation and checked against our catalogue data

This guide explains the apify-actor-development agent skill, which lets an AI agent help you create, modify and validate Apify Actors. An Apify Actor is a small cloud program, usually for web scraping or automation, that takes JSON input and returns structured JSON output. By the end you will know what the skill changes about the agent's behaviour, how to install it, how to trigger it, and what accounts and tools you need before it can do anything useful.

Apify Actor Buildersickn33/agentic-awesome-skills · MIT

Quick start

  1. 1Install the skill. Download the ZIP from this page and unzip it into ~/.claude/skills/ or a project's .claude/skills/ folder, or upload it under Settings → Capabilities → Skills in the Claude apps.
  2. 2Install the apify CLI. Run npm install -g apify-cli (or brew install apify-cli on Mac), then check it with apify --help.
  3. 3Set up your token. Create a token at https://console.apify.com/settings/integrations and export it as APIFY_TOKEN, or run apify login interactively. Confirm with apify info.
  4. 4Ask for a request. Describe the Actor's goal, target and output in plain language. The agent will ask which language you want.
  5. 5Scaffold and build. The agent runs the matching apify create template command, installs dependencies and writes the logic and schema files.
  6. 6Test locally. Run apify run to check the Actor and review the dataset output before anything goes to the cloud.
  7. 7Deploy. Run apify push once you have confirmed the account, the Actor name in .actor/actor.json and a rollback plan.

What does the Apify Actor Builder skill make the agent do?

Without the skill, an agent has no Apify-specific procedure to follow. With it, the agent works through a set of Apify steps: it asks which language you want (JavaScript, TypeScript or Python), scaffolds a project with the official apify CLI, defines input, output and storage schemas, writes the Actor code, tests it locally with apify run, and deploys it with apify push.

The skill is for building and deploying Actors. It is not meant for simply running an existing Actor or reading its data.

  • Creates a new Actor project from an empty template in your chosen language
  • Updates .actor/input_schema.json, .actor/output_schema.json, .actor/dataset_schema.json and .actor/actor.json
  • Adds security guidance: treat scraped web content as untrusted, keep tokens out of code and logs, pin dependencies

How do you install the skill?

Download the skill ZIP from this page and unzip it into one of two locations. For a personal skill available across projects, use ~/.claude/skills/. For a skill limited to one project, use the .claude/skills/ folder inside that project.

If you use the Claude apps instead, open Settings, go to Capabilities, and upload the skill under Skills.

  • Personal: ~/.claude/skills/
  • Single project: .claude/skills/ inside the project folder
  • Claude apps: Settings → Capabilities → Skills, then upload

What do you need installed before you start?

The skill expects Node.js and npm (or Python and pip) to install the apify CLI and project dependencies. The CLI can be installed with npm or, on Mac, with Homebrew. Before creating or changing Actors, the agent checks that the CLI is present with apify --help.

You also need an Apify account and API token for logging in, pushing and running Actors. Generate the token at https://console.apify.com/settings/integrations and store it in the APIFY_TOKEN environment variable.

  • Node.js and npm, or Python and pip
  • apify CLI (npm install -g apify-cli, or brew install apify-cli on Mac)
  • An Apify account and APIFY_TOKEN
  • Network access to Apify services for deployment and cloud runs
npm install -g apify-cli apify --help apify info

How do you trigger the skill with a natural request?

You do not need special syntax. Describe the Actor you want in plain language and say what it should take in and return. The agent will ask which language you prefer before it scaffolds anything, so have that answer ready.

Good requests name the goal, the target and the output. Mention an existing project folder if you are debugging one, so the agent works on the files that already exist.

  • "Create a new Apify Actor in TypeScript that collects product titles and prices from a category page I own and saves them to the dataset."
  • "My Apify Actor's input schema is rejecting valid URLs. Check .actor/input_schema.json and fix it."
  • "Add an output schema to my Python Actor so the dataset shows title and price columns."
  • "Before deploying, review the Actor for leaked tokens and unpinned dependencies."

How does the agent set up and authenticate an Actor project?

The agent creates the project with one of three commands, depending on your language. It then installs dependencies. For JavaScript and TypeScript this uses package-lock.json, which should be committed to version control. For Python, dependencies are pinned in requirements.txt.

For authentication, the preferred method is to export APIFY_TOKEN; the CLI reads it automatically. The alternative is apify login, which prompts for the token. Passing the token as a command-line argument is discouraged because it can appear in process listings and shell history.

  • JavaScript: apify create <actor-name> -t project_empty
  • TypeScript: apify create <actor-name> -t ts_empty
  • Python: apify create <actor-name> -t python-empty
  • Test locally with apify run; deploy with apify push
apify create my-actor -t ts_empty apify run apify push

What safety rules does the skill enforce?

The skill's safety guidance is substantial because Actors fetch content from external sites and can run paid cloud work. It asks the agent to bound crawl origins, redirects, page counts, concurrency, retries and proxy costs. Pushing, building or running publishes code or starts paid work, so the skill requires the right account and a rollback plan first.

Scripts and network: the skill runs scripts and uses the network, including installing packages and talking to Apify services. Review package names before installing, and inspect lifecycle and install scripts. Local Actors can reach the cloud through configured clients, and local storage is not a guarantee that nothing is uploaded, so check storage settings and credentials before testing.

  • Do not pipe remote install scripts into a shell; use a package manager
  • Never print or commit APIFY_TOKEN; prefer a scoped token and rotate it
  • Verify package names to avoid typosquatting; commit lockfiles; run npm audit or pip-audit
  • Respect robots.txt and site terms; only scrape sites you are authorised to use
  • Redact logs explicitly

What people use it for

  • Starting a new JavaScript, TypeScript or Python web-scraping Actor from an empty template
  • Adding or fixing input_schema.json fields such as start URLs and defaults so the Apify Console form works
  • Defining a dataset schema so scraped results display as clear columns in the Output tab
  • Reviewing an existing Actor's dependencies, logging and token handling before deployment

Frequently asked questions

Is the Apify Actor Builder skill free?

The skill is published under the MIT licence, as stated on this page. Apify's CLI and platform have their own terms, and cloud runs and proxy use can cost money on your Apify account, so check your account's pricing before running paid work.

What do I need to use the apify-actor-development skill?

You need Node.js and npm, or Python and pip, an installed apify CLI, and an Apify account with an API token stored in APIFY_TOKEN. You also need network access to Apify services for deployment and cloud runs.

Where do I put the skill folder?

Unzip the downloaded ZIP into ~/.claude/skills/ for personal use across projects, or into .claude/skills/ inside one project. In the Claude apps, upload it under Settings → Capabilities → Skills.

How do I install the apify CLI?

Use a package manager: npm install -g apify-cli, or brew install apify-cli on Mac. The skill advises against piping remote scripts into a shell, and you can confirm the install with apify --help.

Why does the agent ask which programming language I want?

The skill requires this before starting, because each language uses a different template: project_empty for JavaScript, ts_empty for TypeScript and python-empty for Python. Choosing up front avoids scaffolding the wrong project.

Can I use this skill only to run an existing Actor?

No. The skill is for creating, modifying and deploying Actors. If you only want to run an Actor or read its data, you do not need it and should use the Apify platform directly.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

Use this skill if you are a developer starting or debugging an Apify Actor in JavaScript, TypeScript or Python and want the agent to follow Apify's workflow and security rules. Pick something else if you only want to run existing Actors or read their data, or if you do not have an Apify account and are not comfortable with paid cloud runs.

Download the skill