轻量研究
lite-research
by chand1012
Perform lite research using web search and web fetch tools. Use when user asks to "research" something (not "deep research"). Uses advanced search strategies with synonyms, Boolean operators, and alternative terminology. Never spawns sub-agents.
安装
claude skill add --url github.com/openclaw/skills/tree/main/skills/chand1012/lite-research文档
Lite Research Skill
This skill performs quick research by searching the web and reading the top results using built-in tools with advanced search strategies.
When to Use This Skill
Use this skill when the user says "research <topic>" (but NOT "deep research"). This skill:
- Uses web_search to find results for multiple related queries using search best practices
- Uses web_fetch to read the top 5 results for each query
- Does NOT spawn sub-agents
- Provides a summary of findings with sources
Advanced Search Strategy
Instead of basic queries, this skill uses research-proven techniques and limits queries to maximum 5:
1. Query Variations with Synonyms and Alternatives
For a topic like "artificial intelligence", create 2-5 queries like:
- "artificial intelligence" OR "AI" OR "machine learning"
- "artificial intelligence" overview introduction basics
- "artificial intelligence" recent developments 2024 2025
- "artificial intelligence" challenges problems issues
- "artificial intelligence" applications uses examples
Limit: Maximum 5 queries total to balance thoroughness with efficiency.
2. Boolean Operators and Search Techniques
- Use OR for synonyms: "climate change" OR "global warming"
- Use quotation marks for exact phrases
- Include both singular and plural forms
- Add time constraints when relevant
3. Alternative Terminology
Consider:
- Different spellings (globalization vs globalisation)
- Technical vs common terms ("myocardial infarction" vs "heart attack")
- Regional variations ("university" vs "college")
- Historical terminology changes
4. Multi-Perspective Approach
Create queries from different angles:
- Basic overview: "{topic} overview introduction"
- Current state: "{topic} 2024 2025 recent developments"
- Critical analysis: "{topic} challenges problems issues"
- Applications: "{topic} applications uses examples"
Implementation Workflow
When user says "research <topic>":
-
Analyze the topic - identify key concepts and potential synonyms
-
Create 2-5 strategic queries maximum using:
- Synonyms with OR operators
- Different query types (overview, recent, critical, applications)
- Time-specific terms when relevant
- Quotation marks for exact phrases
- Important: Never exceed 5 distinct queries
-
Execute searches with count=5 for each query (max 5 queries total)
-
Fetch and analyze content from results
-
Compile findings focusing on:
- Key concepts and definitions
- Current developments and trends
- Different perspectives and applications
- Notable challenges or controversies
Example Queries for "renewable energy"
- "renewable energy" OR "clean energy" OR "green energy"
- "renewable energy" overview introduction "current state"
- "renewable energy" 2024 2025 "recent developments" trends
- "renewable energy" challenges problems limitations issues
- "renewable energy" applications examples projects case studies
Maximum 5 queries total - adjust based on topic complexity and available information.
Important Notes
- Only triggers on "research" commands, not "deep research"
- Never spawn sub-agents - use direct tool calls
- Focus on comprehensive but quick overview
- Provide sources and URLs for transparency
- Adapt query strategy based on topic complexity
相关 MCP 服务
Puppeteer 浏览器控制
编辑精选by Anthropic
Puppeteer 是让 Claude 自动操作浏览器进行网页抓取和测试的 MCP 服务器。
✎ 这个服务器解决了手动编写 Puppeteer 脚本的繁琐问题,适合需要自动化网页交互的开发者,比如抓取动态内容或做端到端测试。不过,作为参考实现,它可能缺少生产级的安全防护,建议在可控环境中使用。
Brave 搜索
编辑精选by Anthropic
Brave Search 是让 Claude 直接调用 Brave 搜索 API 获取实时网络信息的 MCP 服务器。
✎ 如果你想让 AI 助手帮你搜索最新资讯或技术文档,这个工具能绕过传统搜索的限制,直接返回结构化数据。特别适合需要实时信息的开发者,比如查 API 更新或竞品动态。不过它依赖 Brave 的 API 配额,高频使用可能受限。
网页抓取
编辑精选by Anthropic
Fetch 是 MCP 官方参考服务器,让 AI 能抓取网页并转为 Markdown 格式。
✎ 这个服务器解决了 AI 直接处理网页内容时格式混乱的问题,适合需要让 Claude 分析在线文档或新闻的开发者。不过作为参考实现,它缺乏生产级的安全配置,你得自己处理反爬虫和隐私风险。