multi-search-engine

Multi search engine integration with 17 engines (8 CN + 9 Global). Supports advanced search operators, time filters, site search, privacy engines, and Wolfra...

Security Scan

The skill's documentation matches a multi-engine search tool, but metadata and packaging inconsistencies plus broad web-scraping instructions leave unanswered questions you should verify before installing.

Purpose & Capability

The name/description (multi search across 17 engines) aligns with the provided SKILL.md, config.json, and reference docs. However package metadata is inconsistent: package.json lists main="multi-search.py" but no code files are present, and file-level metadata (_meta.json, CHANGELOG) show a different version (2.0.1) than the registry metadata (1.0.0). These mismatches suggest the published bundle may be incomplete or not the intended release.

Instruction Scope

SKILL.md and references instruct the agent to perform web fetches and HTML scraping across many third-party search endpoints and include examples using powerful search operators (e.g., site:, intext:, filetype:). While expected for a search tool, these instructions implicitly direct the agent to fetch arbitrary external URLs and to perform searches that could surface sensitive information (e.g., searching for 'intext:password filetype:txt'). The docs also mention scraping Chinese engines without API keys but give no guidance about throttling, robots.txt, or legal/TOS considerations.

Install Mechanism

This is an instruction-only skill with no install spec and no code files—nothing will be written or installed by the skill package itself. That minimizes install-time risk, but it also means the runtime behavior depends entirely on the agent executing the documented web requests/parsing logic (which isn't provided as packaged code).

Credentials

The skill declares no required environment variables. SKILL.md lists optional API keys (Google, Bing, Wolfram) which are reasonable for improving API usage. This is proportionate. However, because there is no packaged code, it's unclear where optional .env keys would be used; verify how/where they'd be read if you install/run a runtime implementation.

Persistence & Privilege

The skill does not request 'always: true' and is user-invocable with normal autonomous invocation allowed. It does not declare config-path or system modifications. No persistence or elevated privileges are requested in the bundle.

What to consider before installing

This package appears to be documentation for a multi-engine search helper, but there are packaging inconsistencies and missing implementation files. Before installing or enabling it: (1) ask the publisher to provide the actual runtime code (multi-search.py or equivalent) and verify the code matches the docs; (2) confirm the correct owner/version and that metadata mismatches are intentional; (3) review the runtime implementation for any code that fetches arbitrary URLs, logs or forwards results, or reads environment files; (4) avoid supplying API keys until you verify how they're stored/used; (5) be aware that the skill's scraping behavior can surface sensitive data and may violate search engines' terms—prefer to run it in a restricted environment or request a vetted implementation with rate-limiting and robots.txt respect. If the publisher cannot supply clear, matching source code and provenance, treat the package as untrusted.

Multiple search engine integration with support for Chinese and global search engines.

Supported Search Engines

Chinese Search Engines (8 total)

• 🔍 Baidu
• 🦉 360 Search
• 🔍 Bing China
• 🇨🇳 Sogou
• 🥇 Weibo
• 🧭 Baidu Scholar
• 🌐 Zhihu
• 🔖 Bing Scholar

Global Search Engines (9 total)

• 🌍 Google
• 🔍 Bing
• 🦆 DuckDuckGo
• 🐙 Brave Search
• 🌐 Startpage (privacy-focused)
• 🔬 Google Scholar
• 🧠 WolframAlpha (knowledge queries)
• 📰 Brave News
• 🍴 Hacker News

When to Use

• User asks to search the web for current information
• Need to find recent news
• Search for academic papers
• Ask knowledge-based computational questions (WolframAlpha)
• Prefer privacy-focused search
• Chinese language/region-specific search

How to use

Basic Search

from multi_search import search

results = search(
    query="artificial intelligence latest developments", 
    count=5, 
    engine="google" # optional, default searches all configured
)

for result in results:
    print(result.title, result.url, result.snippet)

Command Line

# Search all engines
multi-search "what is openclaw" --count 10

# Search specific engine
multi-search "machine learning" --engine bing

# Search with time filter
multi-search "AI news" --days 7

# Site-specific search
multi-search "tavily" --site github.com

# WolframAlpha computational knowledge
multi-search "integral of x^2 sinx dx" --engine wolframalpha

Configuration

Configure API Keys (optional)

Create .env file in skill directory:

# Optional API keys (most engines work without API keys for basic search)
GOOGLE_API_KEY=your_key
BING_API_KEY=your_key
WOLFRAM_APP_ID=your_id
WOLFRAM_API_KEY=your_key

# Enable/disable specific engines
ENABLED_ENGINES=baidu,bing,google,duckduckgo,bing,startpage,brave

**Most Chinese engines work without API keys via web scraping.

Response Format

[
  {
    "title": "Result Title",
    "url": "https://example.com",
    "snippet": "Text summary",
    "engine": "google",
    "position": 1
  }
]

Features

• Advanced search operators: support for (site:, inurl:, filetype, exact phrase
• Time filtering: search within N days/months/years
• Language preference: auto-language detection based on query language
• Fallback: if one engine fails, falls back to next
• No API key required: most engines work without API keys via public endpoints

Dependencies

pip install requests beautifulsoup4

## Credits

- Inspired by [searchapi](https://github.com/tobias neutralone