什么是 Exa Researcher？

Exa Researcher 是一个全面的研究套件，旨在赋予 AI 智能体执行高保真网页搜索、系统化抓取和结构化数据提取的能力。通过利用 Exa AI 先进的模型上下文协议 (MCP) 工具，该技能允许智能体超越表面层面的关键词匹配，对公司、人员和技术代码库进行深层语义搜索和背景调查。

作为 Openclaw Skills 库的核心组件，Exa Researcher 弥补了静态 LLM 知识与网络上实时动态数据之间的差距。它专为需要其智能体以高精度和高相关性执行专业级市场研究、竞争分析和技术文档综合的开发者而构建。

下载入口:https://github.com/openclaw/skills/tree/main/skills/sy2ruto/exa-researcher

安装与下载

1. ClawHub CLI

从源直接安装技能的最快方式。

2. 手动安装

将技能文件夹复制到以下位置之一

优先级：工作区 > 本地 > 内置

3. 提示词安装

将此提示词复制到 OpenClaw 即可自动安装。

Exa Researcher 应用场景

• 进行竞争情报收集，获取详细的公司概况以进行市场分析。
• 通过针对性的人员搜索进行专业背景研究和专家发现。
• 从文档和存储库中提取技术代码上下文和 API 实现模式。
• 使用高级搜索过滤器自动进行行业新闻监控和趋势分析。
• 执行需要异步处理和进度跟踪的多源深度研究项目。

1. AI 智能体识别特定的研究目标，例如查找公司的领导层或 API 的使用模式。
2. 技能根据查询类型选择优化的 Exa 工具（例如 company_research_exa 或 get_code_context_exa）。
3. 语义搜索或抓取请求被发送到 Exa API，利用日期范围、域名或地区等过滤器来缩小范围。
4. 技能处理原始响应，提取高亮内容、片段和结构化元数据供智能体使用。
5. 对于长时间运行的任务，深度研究工作流会维持状态，允许智能体检查进度并随时间汇总发现。
6. 最终综合的数据返回到智能体的上下文中，以供进一步行动或生成报告。

Exa Researcher 配置指南

要将 Exa Researcher 集成到您的环境中，请确保已设置 API 凭据并将该技能添加到您的智能体中。

# 设置您的 Exa AI API 密钥
export EXA_API_KEY='your_api_key_here'

# 通过 Openclaw Skills 管理器安装技能
openclaw add skill exa-researcher

Exa Researcher 数据架构与分类体系

该技能将研究数据组织成结构化格式，以确保 AI 智能体可以轻松解析和综合信息。

name: exa-researcher
description: "Comprehensive research capabilities using Exa AI MCP tools. Use when you need to conduct web research, gather company information, find people, extract code context, or perform deep research tasks. Specifically triggers for: web searches, company research, people search, code context extraction, web crawling, and deep researcher workflows. Perfect for market research, competitive analysis, technical research, news gathering, and any knowledge-intensive research needs."

Exa Researcher

Overview

Exa Researcher provides powerful research capabilities through Exa AI's MCP tools, including web searches (basic and advanced), company and people research, code context extraction, web crawling, and deep research workflows.

Core Capabilities

1. Web Search (Basic & Advanced)

Basic Web Search (web_search_exa)

• Quick web searches for general information
• Ideal for: factual queries, news, quick facts

Advanced Web Search (web_search_advanced_exa)

• More sophisticated search with filters and ranking
• Ideal for: detailed research, finding specific content types, narrowing results by freshness, region, etc.

2. Company Research (company_research_exa)

• Gather comprehensive information about companies
• Ideal for: competitive analysis, partner research, due diligence
• Use cases: company overview, financial data, recent news, leadership, products/services

3. People Search (people_search_exa)

• Find information about people by name or query
• Ideal for: background research, networking, finding experts
• Use cases: professional backgrounds, expertise areas, recent activity

4. Code Context (get_code_context_exa)

• Extract code context and technical information
• Ideal for: technical research, understanding codebases, finding code examples
• Use cases: understanding APIs, finding implementation patterns, technical documentation research

5. Web Crawling (crawling_exa)

• Systematic web crawling for comprehensive information gathering
• Ideal for: in-depth site research, content extraction, archiving
• Use cases: website analysis, content aggregation, research from specific domains

6. Deep Research (deep_researcher_start, deep_researcher_check)

• Long-running research workflows with progress tracking
• Ideal for: comprehensive multi-source research, trend analysis, information synthesis
• Use cases: research papers, market reports, feasibility studies, competitive intelligence

Usage Guidelines

When to Use Each Tool

Research Workflow

1. Identify the research goal - What information do you need?
2. Select the appropriate tool - Match the task to the tool above
3. Execute the search/research - Call the Exa AI tool
4. Analyze results - Synthesize findings
5. Iterate if needed - Refine queries or use additional tools

Advanced Use Cases

Competitive Analysis

• Use company_research_exa for multiple competitors
• Combine with web_search_advanced_exa for recent news
• Document findings in structured format

Market Research

• web_search_advanced_exa for market trends
• crawling_exa for industry reports and whitepapers
• Synthesize findings into market overview

Technical Documentation Research

• get_code_context_exa for API documentation
• Find code examples and implementation patterns
• Extract technical specifications

Background Research (People/Companies)

• people_search_exa for expertise matching
• company_research_exa for organizational context
• Combine for comprehensive profiles

Long-Form Research Projects

• Use deep_researcher_start to initiate workflow
• Monitor progress with deep_researcher_check
• Let it run for complex, multi-source research needs

Tips for Best Results

• Be specific with queries - More precise queries yield better results
• Combine tools - Multiple tools can provide comprehensive coverage
• Use filters wisely - Especially with advanced search (date ranges, regions, etc.)
• Iterate - Refine queries based on initial results
• Store important findings - Save significant research to memory
• Consider search volume - Complex searches may take longer