Skip to content
Back to skills

Web Search Liliang Cn

ASecurity

Perform comprehensive web searches and research using the MCP Web Search Server.

  • 57 stars
  • 0 votes
  • 0 copies
  • 9 views
  • Added May 31, 2026
ai-agentsgodocumentation

Works with

  • mcp

Security analysis

A100/100

Scanned May 31, 2026

npx -y skills add diegosouzapw/awesome-omni-skill --skill web-search-liliang-cn --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Web Search Liliang Cn?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Web Search Liliang Cn
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/diegosouzapw-web-search-liliang-cn/badge)](https://www.skillsdirectory.com/skills/diegosouzapw-web-search-liliang-cn)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: web-search
description: Perform comprehensive web searches and research using the MCP Web Search Server.
user-invocable: true
---

# Web Search Assistant

You have access to powerful web search tools via the `mcp-websearch-server`. Use these tools to answer user queries with up-to-date information, perform deep research, and verify facts.

## Available Tools

1.  **`websearch_basic`**: 
    -   **Use for**: Quick lookups, fact-checking, or finding a specific URL.
    -   **Returns**: Titles, URLs, and short snippets.
    -   **Cost**: Low latency.

2.  **`websearch_with_content`**: 
    -   **Use for**: Deep dives into specific pages or when snippets are insufficient.
    -   **Returns**: Full extracted text content from the page (intelligent extraction).
    -   **Cost**: Medium latency (fetches page).

3.  **`websearch_multi_engine`**: 
    -   **Use for**: Broad research on complex or controversial topics.
    -   **Returns**: Aggregated results from Bing, Brave, and DuckDuckGo.
    -   **Cost**: Higher latency (multiple searches).

4.  **`websearch_ai_summary`**: 
    -   **Use for**: Getting a pre-digested, AI-ready summary of a topic.
    -   **Returns**: Structured markdown optimized for LLM analysis.

## Recommended Workflows

### 🔍 Deep Research
When the user asks for in-depth research on a topic:
1.  **Broad Sweep**: Use `websearch_multi_engine` with a query like "overview of [topic]" or specific sub-questions.
2.  **Identify Sources**: Look at the results. If a result looks promising but the snippet is cut off, use `websearch_with_content` on that specific URL (or use the tool if it supports URL input, otherwise search specifically for that page).
3.  **Synthesize**: Combine information from multiple sources. Cite your sources using the URLs provided.

### ✅ Fact Checking
When the user asks "Is [X] true?":
1.  **Quick Check**: Use `websearch_basic` with a precise query.
2.  **Verify**: If results are conflicting, use `websearch_multi_engine` to see different perspectives.

### 📰 Latest News
When the user asks for recent events:
1.  Use `websearch_basic` or `websearch_multi_engine` with "latest news [topic]".
2.  Prioritize results with recent dates in the snippet (if visible) or title.

## Tips
-   Always cite the source (URL) when providing information found via search.
-   If the first search yields no results, try a broader keyword or a different search tool (e.g., switch from `websearch_basic` to `websearch_multi_engine`).
-   For technical documentation or coding questions, prefer `websearch_with_content` to get code examples and full context.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.