Crawl

Crawl Skill

Required Environment Variables

• None required.

Invocation Contract

• Call run_skill_script with:
• skill_name: crawl
• script_name: one of fetch_page, fetch_js_page, extract_links, extract_tables, crawl_site, check_robots
• Put required positional parameters in args.
• Put optional value flags in kwargs using the keys listed below.
• For flag-only boolean options, use args (not kwargs).

Scripts API

script_name: fetch_page

• args: [""]
• optional kwargs: timeout, output

script_name: fetch_js_page

• args: [""]
• optional kwargs: wait, selector, timeout, output
• optional boolean flag --html: pass via args as "--html"

script_name: extract_links

• args: [""]
• optional kwargs: filter, timeout, output
• --absolute is enabled by default.

script_name: extract_tables

• args: [""]
• optional kwargs: format, timeout, output
• format: json | csv

script_name: crawl_site

• args: [""]
• optional kwargs: max_pages, max_depth, timeout, output
• --same-domain is enabled by default.

script_name: check_robots

• args: [""]
• optional kwargs: user_agent

run_skill_script Examples

• JS-rendered fetch with selector:

{
  "skill_name": "crawl",
  "script_name": "fetch_js_page",
  "args": ["https://example.com"],
  "kwargs": {
    "wait": 1200,
    "selector": "main",
    "timeout": 45000
  }
}

• Include raw HTML output:

{
  "skill_name": "crawl",
  "script_name": "fetch_js_page",
  "args": ["https://example.com", "--html"]
}

Notes:

• fetch_js_page requires Playwright + Chromium in the execution environment.

Direct CLI Examples

python agent_skills/skills/crawl/scripts/fetch_page.py https://example.com --timeout 20
python agent_skills/skills/crawl/scripts/fetch_js_page.py https://example.com --wait 1200 --selector main
python agent_skills/skills/crawl/scripts/extract_links.py https://example.com --filter /docs
python agent_skills/skills/crawl/scripts/extract_tables.py https://example.com --format csv --output tables.csv
python agent_skills/skills/crawl/scripts/crawl_site.py https://example.com --max-pages 20 --max-depth 2 --output crawl.json
python agent_skills/skills/crawl/scripts/check_robots.py https://example.com --user-agent "DataBot/1.0"

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: datalayer
• Source: datalayer/agent-skills
• License: BSD-3-Clause
• Homepage: https://agent-skills.datalayer.tech

Install and usage instructions live in the source repository linked above.