Easy Search

Simple web search using multiple search engines with no API key required. Supports Google, Bing, DuckDuckGo, Baidu, Sogou, 360, Brave, Yandex. Features time...

Security Scan

The skill's code and instructions align with its description: it performs HTML-based web searches across multiple engines, uses optional duckduckgo-search/requests, respects proxy env vars, and caches results locally — nothing requested is disproportionate to that purpose.

Purpose & Capability

The name/description (multi-engine web search without API keys) matches the included script and SKILL.md. The code implements HTML scraping, optional duckduckgo_search usage, engine selection, snippet extraction, time filters, proxy support, and local caching — all of which are expected for this functionality.

Instruction Scope

Runtime instructions tell the agent to run scripts/search.py and to set proxy env vars if needed. The code performs network requests to many search engines, can run interactively and perform diagnostics, and writes cache files under the user's home (~/.easysearch_cache). The SKILL.md does not declare the cache path explicitly; aside from local caching and contacting search engines, there are no instructions to read unrelated system files or exfiltrate data.

Install Mechanism

There is no install spec and no external downloads; this is effectively an instruction+script bundle. That is low-risk from an install-supply perspective (nothing is being fetched or extracted at install time).

Credentials

The skill declares no required credentials or special environment variables. It does honor common proxy variables (ALL_PROXY, HTTPS_PROXY, etc.), which is appropriate for a network tool. No unrelated secrets or external credentials are requested.

Persistence & Privilege

The skill is not forced-always, does not ask to modify other skills, and only writes to its own cache directory in the user's home. It does not request elevated privileges or system-wide configuration changes.

Assessment

This skill appears to do what it claims: it scrapes public search engine HTML, makes network requests, and stores a small cache in ~/.easysearch_cache. Before installing, consider: (1) don't use it to search for sensitive secrets because queries go to external search engines; (2) scraping may be brittle or blocked by some engines and could raise TOS concerns; (3) the tool respects proxy env vars — if you use a corporate or local proxy, your queries will go through it; (4) if you want extra assurance, review the full scripts locally or run them in a sandboxed/non-privileged account. Overall the skill is internally consistent and does not request disproportionate credentials or system access.

Easy Search Skill

A powerful web search skill that doesn't require API keys. It uses direct HTTP requests and the duckduckgo_search library for reliable results.

Features

• No API key required - Uses public search interfaces
• 9 search engines - Google, Bing, DuckDuckGo, Baidu, Sogou, 360, Brave, Yandex, Startpage
• Auto engine selection - Automatically picks the best available engine
• Snippet extraction - Get result previews, not just titles
• Time filtering - Filter by day/week/month/year
• Interactive mode - Continuous searching with live engine switching
• Network diagnostics - Built-in connectivity checking and proxy hints
• Proxy support - Respects ALL_PROXY, HTTPS_PROXY environment variables
• Result caching - Faster repeated searches
• Smart failover - Auto-switches to backup engines on failure

Requirements

• Python 3.6+
• Optional: pip install duckduckgo-search requests for better results

Commands

Basic Search

# Simple search (positional argument)
python3 {baseDir}/scripts/search.py Python tutorial

# With flag
python3 {baseDir}/scripts/search.py --query "your search terms"

# Specify engine
python3 {baseDir}/scripts/search.py "AI news" --engine bing

# More results
python3 {baseDir}/scripts/search.py "React hooks" --results 10

# JSON output (default)
python3 {baseDir}/scripts/search.py "Vue.js" --format json

# Markdown output
python3 {baseDir}/scripts/search.py "machine learning" --format md

# Simple output (quick reading)
python3 {baseDir}/scripts/search.py "Docker" --format simple

# CSV output
python3 {baseDir}/scripts/search.py "Kubernetes" --format csv

Time Filtering

# News from last 24 hours
python3 {baseDir}/scripts/search.py "AI news" --time day

# Results from last week
python3 {baseDir}/scripts/search.py "React 19" --time week

# Results from last month
python3 {baseDir}/scripts/search.py "Python 3.13" --time month

# Results from last year
python3 {baseDir}/scripts/search.py "best laptops 2024" --time year

Interactive Mode

# Start interactive mode
python3 {baseDir}/scripts/search.py --interactive

# In interactive mode:
# [auto]> Python tutorial          # Search with auto engine
# [auto]> :engine duckduckgo       # Change engine
# [duckduckgo]> React hooks        # Search with new engine
# [duckduckgo]> :format simple     # Change output format
# [duckduckgo]> :check             # Check engine health
# [duckduckgo]> :quit              # Exit

Network Diagnostics

# Run network diagnostics
python3 {baseDir}/scripts/search.py --diagnose

# Check which engines are available
python3 {baseDir}/scripts/search.py --check-engines

Engine Management

# Clear cache
python3 {baseDir}/scripts/search.py --clear-cache

# Disable auto-fallback (stay on specified engine)
python3 {baseDir}/scripts/search.py "query" --engine google --no-fallback

Search Engines

Engine	Best For	Notes
auto	General use	Auto-selects best available engine
startpage	Privacy, global	Privacy-focused, reliable HTML parsing
duckduckgo	Privacy, global	Preferred, uses library if installed
bing	Global search	Good reliability, decent snippets
google	Best results	May need proxy in some regions
baidu	Chinese content	Strong anti-crawl measures
sogou	Chinese content	Good alternative to Baidu
so360	Chinese content	Another Chinese option
brave	Privacy-focused	Growing index
yandex	Russian/Global	Works in some regions

Output Formats

JSON

{
  "query": "search terms",
  "engine": "bing",
  "method": "HTML scraping",
  "results": [
    {
      "title": "Result Title",
      "url": "https://example.com",
      "snippet": "Brief description..."
    }
  ],
  "total": 5
}

Markdown

## Search Results: your terms

Engine: **BING**
Method: _HTML scraping_

1. [Result Title](https://example.com)
   > Brief description...

2. [Another Title](https://another.com)
   > More info...

Simple

1. Result Title
   https://example.com
   Brief description...

2. Another Title
   https://another.com
   More info...

Examples

# Quick search for documentation
python3 {baseDir}/scripts/search.py "Python requests library docs"

# Search Chinese content with Sogou
python3 {baseDir}/scripts/search.py "人工智能最新进展" --engine sogou

# Get recent news in markdown format
python3 {baseDir}/scripts/search.py "OpenAI news" --time day --format md

# Use interactive mode for multiple searches
python3 {baseDir}/scripts/search.py -i

# Search with time filter
python3 {baseDir}/scripts/search.py "React 19 features" --time month --results 10

Proxy Configuration

Set environment variables for proxy support:

# Set proxy
export ALL_PROXY="http://127.0.0.1:7890"
export HTTPS_PROXY="http://127.0.0.1:7890"

# Then run search
python3 {baseDir}/scripts/search.py "your query"

Troubleshooting

Problem	Solution
Empty results	Try a different engine with --engine
Connection timeout	Check proxy settings or use --timeout 60
Rate limiting	Wait between searches or switch engines
DuckDuckGo slow	Install library: pip install duckduckgo-search
No snippets	Some engines have limited snippet extraction
Google blocked	Use proxy or switch to Bing/DuckDuckGo
All engines fail	Run --diagnose to check network

Tips for Best Results

1. Use auto engine - Let the tool pick the best available engine
2. Install dependencies - pip install duckduckgo-search requests for better reliability
3. Use time filters - For news and recent topics, use --time day or --time week
4. Try Chinese engines - Use sogou or so360 for Chinese content
5. Interactive mode - Great for multiple related searches
6. Cache is your friend - Repeated searches are instant (1 hour cache)
7. Run diagnostics - Use --diagnose to troubleshoot network issues

Version History

• v1.1.0 - Added Startpage engine with proper parsing, fixed HTML entity decoding in URLs
• v1.0.1 - Added network diagnostics, better proxy hints, improved error messages
• v1.0.0 - Added 4 new engines (Sogou, 360, Brave, Yandex), time filtering, interactive mode, snippet extraction, engine health caching
• v4.0.0 - Integrated duckduckgo_search library, auto engine selection
• v1.0.0 - Initial release with basic search