How to use the Twitter/X Scraping with twscrape agent skill

What the twscrape skill makes the agent do

An agent skill is a folder of instructions and reference files that Claude loads when a task matches it. The twscrape skill is focused on one tool: the twscrape Python library, which calls Twitter/X's internal GraphQL web API rather than the official paid API.

With the skill loaded, the agent knows how to add logged-in accounts, rotate between them when rate limits are hit, and configure proxies. It can then write scripts that search tweets, look up user profiles, collect followers or replies, and fetch trends, or run the twscrape CLI for you.

• Search tweets by query, with optional Top-tweet ordering
• Look up users by username or ID and read profile counts
• Collect followers, following, tweet replies, retweeters and user media
• Fetch trending topics by category such as news or sport
• Manage accounts and proxies through Python or the CLI

Install the twscrape skill

Download the ZIP for the skill from this page and unzip it. The folder contains SKILL.md, plus a references folder with installation.md, examples.md and index.md.

You can install it in one of two ways. For your personal Claude setup, unzip it into ~/.claude/skills/. For a single project, unzip it into that project's .claude/skills/ folder. Alternatively, in the Claude apps open Settings, then Capabilities, then Skills, and upload the skill there.

Set up the Python library the skill depends on

The skill needs Python with pip and the twscrape package. The library's own install instructions use the following command, and the skill's reference files repeat it.

The source lists Python 3.7 or newer, asyncio support and an internet connection as requirements. Verification is done with the twscrape --help command and a short Python import check.

• Dependencies such as httpx and aiosqlite are installed automatically
• The default account database is accounts.db in your current working directory

Trigger the skill with natural requests

You do not need special syntax. Describe the data you want and mention Twitter or X. The skill is meant to load when you ask for Twitter/X extraction, monitoring, follower collection or an alternative to the official API.

Be specific about the target and the output. Say what you want to find, how many items, and whether you want a script or a one-off CLI command. The agent will then ask for any missing details it needs, such as which accounts to use.

• Write a Python script that collects the latest 100 tweets about a topic and saves them to JSON
• Get the follower list for this username and count how many have verified status
• Show me how to add two Twitter accounts with cookies and rotate between them
• Run twscrape to fetch the replies to this tweet ID with a limit of 50

Add accounts, proxies and rate-limit handling

The skill explains two ways to add accounts. Cookies are described as the more stable option. Login with username and password is described as less stable. Accounts can also be listed, re-logged-in or retried from the CLI.

Proxies can be set per account, globally in the API, or with the TWS_PROXY environment variable. The source gives the priority order as api.proxy first, then TWS_PROXY, then the account-specific proxy. Rate limits reset every 15 minutes per endpoint, and accounts rotate automatically when limits are reached.

• Use an account file in the form username:password:email:email_password with twscrape add_accounts
• Run twscrape login_accounts, or add --manual for manual email verification
• Check status with twscrape accounts

Safety, scope and limits to know first

This skill runs scripts and makes network requests to Twitter/X. Review any generated code before running it, and only run it with accounts you are allowed to use. The library uses an unofficial interface, so it can break when Twitter changes it, and accounts may be suspended.

The skill is built for bulk collection of profiles and followers. Do not use it to collect private data or other people's personal information without a lawful basis. The source also notes that user_tweets and user_tweets_and_replies return about 3,200 tweets maximum per user, a platform limit.

• Rate limits vary with account age, verification status and activity history
• Email verification needs IMAP access or manual entry, and the default wait is 30 seconds (TWS_WAIT_EMAIL_CODE)