🎁 Get the FREE AI Skills Starter Guide — Subscribe →
BytesAgainBytesAgain
🦀 ClawHub

X Read

by @tylordius

Render and summarize a public X (Twitter) link when you need to read the tweet/article content without logging in.

Versionv1.0.1
Downloads1,780
TERMINAL
clawhub install x-read

📖 About This Skill


name: x-read description: Render and summarize a public X (Twitter) link when you need to read the tweet/article content without logging in.

X-Read Skill

When to use this skill

Use x-read whenever you need to capture the text, media links, and basic metadata from a public X/Twitter permalink without using the API, creating an account, or burning Brave search credits. It works best for thread permalink URLs (https://x.com//status/) and articles posted on X that surface their content on the same page.

Inputs

  • url (required): The exact permalink you want to render. Include the full https://x.com/... path so the Puppeteer browser can navigate directly.
  • What happens when you run it

    1. Puppeteer launches a sandboxed Chromium instance with a realistic user-agent and loads the URL. 2. The script waits for article[data-testid="tweet"] elements, captures the main thread (tweet + up to three replies), and extracts any linked card or media attachments. 3. Media is reprinted as ![](...) markdown with ALT text when available, and the main article body is appended beneath the primary tweet. 4. The output is a short markdown summary listing the thread author, timestamp, text, links, and media so the Telegram chat stays readable.

    Output format

    ## Tweet Thread Summary
    Source: 

    MAIN TWEET by Display Name (@handle)

    *2026-02-21T18:24:00Z* Tweet text …

    Linked Article: Title Media:

  • ![](media1)
  • ![](media2)

  • Replied tweets follow the same pattern with the Reply heading. Media URLs retain the name=large parameter for better quality.

    Limitations and troubleshooting

  • The skill only reads publicly available tweets. If X blocks the navigation with a login wall, you will see an error that logging in is required.
  • Long-form articles sometimes use article.content; the script appends that text to the main tweet so you still get the full written piece.
  • Because the skill uses a browser, it may take a few extra seconds compared to an API call—expect 5–10 seconds per URL.
  • For best results, keep the thread URLs focused and don’t try to feed a feed of multiple unrelated URLs at once; run the tool per link.
  • Testing and safety

  • The skill is read-only: it never posts, likes, DM’s, or authenticates to X.
  • Review the index.js script if X changes its DOM selectors (it relies on data-testid attributes such as tweet, tweetText, User-Name, and card.wrapper).
  • Run the skill manually with openclaw use x-read read_tweet --url before publishing to confirm the output is what you expect.