Smart Web Scraper

Extract structured data from any web page. Supports CSS selectors, auto-detection of tables and lists, JSON/CSV output formats. Use when asked to scrape a we...

This skill appears coherent and implements a normal static-HTML scraper. Before installing or running: (1) Review the full scripts/scraper.py file yourself (the provided view was truncated here) to confirm there are no hidden network callbacks or unexpected behavior; (2) Run it in a sandbox or limited environment first to ensure it only fetches the target sites and does not contact unknown endpoints; (3) Be mindful of legal/terms-of-service and robots.txt — the tool can override robots rules with --ignore-robots; (4) Note that runtime dependency installation (e.g., via `uv run --with` or pip) will fetch code from PyPI — only install packages you trust; (5) Do not supply unrelated credentials (none are required). If you need higher assurance, ask the publisher for a source repository or sign-off and verify the remaining (truncated) portion of the script.

Type: OpenClaw Skill Name: smart-web-scraper Version: 1.0.0 The OpenClaw AgentSkills bundle 'smart-web-scraper' is a web scraping tool that uses standard Python libraries (`urllib`, `BeautifulSoup`, `lxml`) to extract data from user-specified URLs. The `SKILL.md` and `README.md` clearly describe its functionality, and the `scripts/scraper.py` code implements this as described, including features like CSS selectors, table extraction, link analysis, and multi-page crawling. While the script performs network requests to arbitrary URLs and writes output to user-specified file paths, these are core functionalities of a web scraper and do not show evidence of intentional malicious behavior such as data exfiltration to unauthorized endpoints, persistence mechanisms, or arbitrary code execution beyond the skill's stated purpose. The instructions in the markdown files are straightforward and do not contain any prompt injection attempts.