What it does.
Tell the agent a URL and what you want. It returns structured JSON. Handles pagination, JS-rendered pages, rate limits, and the politeness rules.
For when you need a one-off pull and don’t feel like writing puppeteer for the hundredth time.
What’s inside.
Everything you need in the bundle. No tutorials. No assembly.
- 01The SKILL.md
- 02A 2-page guide on writing good "what you want" descriptions
- 03A small library of example scrapes (product pages, listings, directories)
Drop it in. Press go.
- №01
Drop in
Upload to your agent.
- №02
Ask it to scrape
URL + plain-English description of the fields.
- №03
Get JSON back
Or CSV. Or pasted into your spreadsheet, your choice.
How it thinks.
Excerpt · SKILL.md---
name: website-scraper
description: Fetch a URL, extract structured data described in plain English.
allowed_platforms: [openclaw, claude]
tools:
- http.fetch
- browser.render
config:
user_agent: "skillzy-scraper/1.6 (+contact@example.com)"
respect_robots: true
rate_limit_ms: 1200
---
# Website scraper
Input:
- url: the page or list to scrape
- schema: plain-English description of the fields you want
- pagination: optional follow-link rule
Behaviour:
- Render with JS only if needed.
- Respect robots.txt unless explicitly overridden.
- Backoff on 429.
- Return clean JSON conforming to the requested schema.
About the creator.
datajedi@datajedi
Scrapers, parsers, pipelines. Quiet little tools that do the boring work.
- Rating
- 4.6
- Joined
- February 2026
Things people ask.
- Is this legal?
- You’re responsible for what you scrape. The skill respects robots.txt by default.
- Will it bypass logins?
- No. Not a thing we sell.
- JS-heavy sites?
- Yes. It renders when it has to.
$19
Website Scraper