using-web-scraping — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited using-web-scraping (Agent Skill) and scored it 96/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 1 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 1 flagged
The text {match} tells the agent to skip the normal "ask the user first" gate. Used adversarially it removes the human-in-the-loop check before destructive or sensitive actions, turning a normally-gated agent into a fire-and-forget executor.
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
A privacy-minded, agent-facing web-scraping skill that uses headless Chrome (Playwright/Puppeteer) and DuckDuckGo for search. Focuses on: reliable navigation, extracting structured text, obeying robots.txt, and rate-limiting.
/robots.txt before scraping a site.User-Agent strings.login_required instead of attempting to bypass.title, meta description, main text (or best-effort article text), and canonical URL.const { chromium } = require('playwright');
async function ddgSearchAndScrape(query) {
const browser = await chromium.launch({ headless: true });
const page = await browser.newPage({ userAgent: 'open-skills-bot/1.0' });
// DuckDuckGo search
await page.goto('https://duckduckgo.com/');
await page.fill('input[name="q"]', query);
await page.keyboard.press('Enter');
await page.waitForSelector('.result__title a');
// collect top result URL
const href = await page.getAttribute('.result__title a', 'href');
if (!href) { await browser.close(); return []; }
// visit result and extract
await page.goto(href, { waitUntil: 'domcontentloaded' });
const title = await page.title();
const description = await page.locator('meta[name="description"]').getAttribute('content').catch(() => null);
const article = await page.locator('article, main, #content').first().innerText().catch(() => null);
await browser.close();
return [{ url: href, title, description, text: article }];
}
// usage
// ddgSearchAndScrape('open-source agent runtimes').then(console.log);You are an agent with a web-scraping skill. For any `search:` task, use DuckDuckGo to find relevant pages, then open each page in a headless Chrome instance (Playwright/Puppeteer) and extract `title`, `meta description`, `main text`, and `canonical` URL. Always:
- Check and respect robots.txt
- Rate-limit requests (<=1 req/sec)
- Use a clear `User-Agent` and do not execute arbitrary page JS
Return results as JSON: [{url,title,description,text}] or `login_required` if a page needs authentication.npm i playwright and run npx playwright install for browser binaries.pip install playwright and playwright install.page.route to block large assets (images, fonts) when you only need text.~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.