Websearch

Multi-engine web search across 6 engines: DuckDuckGo, DDG Lite, Yahoo, Yahoo JP, Startpage, and Google (headless Chromium). No API keys required. Returns 10+...

Security Scan

The skill's code, instructions, and requirements are consistent with a local multi-engine web-scraping/search aggregator; it doesn't request unrelated credentials or hidden endpoints, but it does use headless browser automation and scraping techniques that have privacy/operational implications.

Purpose & Capability

Name/description describe a multi-engine websearch aggregator and the code/README/SKILL.md implement exactly that: HTTP scraping for several engines plus a Playwright-based headless Google scraper. No unrelated credentials, binaries, or config paths are requested.

Instruction Scope

Runtime instructions and code only perform web requests to listed search engines and local Playwright browser launches. They do not read arbitrary local files or environment variables. Note: queries are sent directly to external search engines (your IP and query will be visible to those sites), and the Google engine contains explicit anti-bot evasion steps (navigator.webdriver override, automation flags).

Install Mechanism

There is no built-in install spec, but SKILL.md requires pip install playwright and running 'playwright install chromium', which downloads browser binaries from Playwright's distribution. This is expected for headless Chromium use; the download comes from the Playwright project rather than an unknown personal server.

Credentials

The skill does not request environment variables, secrets, or unrelated credentials. Its network activity (HTTP fetches, Playwright launching Chromium) is proportional to its stated purpose.

Persistence & Privilege

The skill does not require always:true and does not request modifying other skills or system-wide settings. It will run network requests when invoked; default agent autonomy is allowed by platform but not an intrinsic risk here.

Assessment

This skill appears to do what it says: local scraping of multiple search engines with a Playwright headless Chromium fallback for Google. Before installing, consider: - Privacy: your queries and IP are sent to external search engines; Startpage offers some privacy but other engines will see your requests. - Operational: Playwright will download Chromium (large binary) and the Google scraper uses anti-bot evasion techniques; this may be fragile and could trigger blocking or violate a search engine's terms of service. - Security posture: the Google headless code uses --no-sandbox and navigator.webdriver overrides which are typical for scraping but reduce isolation—avoid running in highly sensitive, privileged environments; prefer a sandboxed VM or container if concerned. - Rate limits: the skill makes parallel requests; avoid heavy automated runs to prevent IP blocking. If these tradeoffs are acceptable, the skill is internally coherent. If you need stronger privacy, run queries through an anonymizing proxy or omit the Google engine.

WebSearch — Multi-Engine Search Aggregator

🔗 GitHub: https://github.com/nirveshdagar/multi-engine-websearch

📦 Git Install:

git clone https://github.com/nirveshdagar/multi-engine-websearch.git

📦 ClawHub Install:

npx clawhub@latest install multi-engine-websearch

---

A free, local web search skill that queries 6 reliable search engines simultaneously (in parallel), merges results, deduplicates URLs, and ranks by cross-engine frequency.

No API keys required. No paid subscriptions. Runs entirely on your machine.

Requirements

• Python 3.9+
• Playwright (pip3 install playwright && python3 -m playwright install chromium)

Playwright is needed for the Google engine (headless Chromium renders JS-based results). All other engines use lightweight HTTP scraping with zero dependencies.

When to Use

Use this skill whenever the user asks you to:

• Search the web for anything
• Find recent news, articles, or information
• Look up a topic, person, product, or event
• Verify a fact with current web results
• Research a topic with multiple sources

Command

python3 ~/.openclaw/workspace/skills/websearch/scripts/search.py "your query" --json

Options

Flag	Default	Description
--json / -j	off	Output as JSON
--num / -n	10	Results per engine
--max / -m	50	Max total results
--engines / -e	all 6	Comma-separated engine list
--list-engines	—	Show available engines

Engines

Engine	Method	Reliability
DuckDuckGo	HTML scrape	⭐⭐⭐⭐⭐
DDG Lite	HTML scrape	⭐⭐⭐⭐
Yahoo	HTML scrape	⭐⭐⭐⭐
Yahoo JP	HTML scrape	⭐⭐⭐⭐
Startpage	HTML scrape	⭐⭐⭐⭐
Google	Headless Chromium	⭐⭐⭐⭐

All engines run in parallel using ThreadPoolExecutor for maximum speed.

Output Format

{
  "query": "your search",
  "total": 35,
  "engines": {"duckduckgo": 10, "google": 8, "yahoo": 10, ...},
  "results": [
    {
      "title": "Result Title",
      "url": "https://example.com",
      "snippet": "Description...",
      "engines": ["duckduckgo", "google", "yahoo"]
    }
  ]
}

Results appearing in more engines are ranked higher (cross-engine consensus).

Core Rules

1. Always use --json flag when parsing results programmatically
2. Summarize top results in natural language for the user
3. Always cite sources with URLs
4. If one engine fails, others compensate — the system is resilient
5. Google engine is slower (~5s) due to headless browser; other engines are fast (~1-2s)