← 返回 Skills 市场
maverick

Maverick Exa

作者 Maverick · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ 安全检测通过
60
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install maverick-exa
功能描述
Exa MCP workflow for web and people research plus hosted MCP server setup using a fixed tool set (`web_search_exa`, `web_search_advanced_exa`, `people_search...
使用说明 (SKILL.md)

Exa MCP Skill

Workflow

  1. Clarify the research goal, scope, and output format.
  2. Select one enabled tool using the guide below: web_search_exa, web_search_advanced_exa, people_search_exa.
  3. Prefer the hosted Exa MCP endpoint with fixed tools in the URL query.
  4. Return concise findings with links and explicit unknowns.
  5. On rate limits (429), reduce query breadth or require an API key.

Tool Policy

  • Enabled now: web_search_exa, web_search_advanced_exa, people_search_exa.
  • Keep all other Exa tools disabled unless the user explicitly asks to expand.

Tool Selection Guide

  • Use web_search_exa for fast, broad discovery when the user asks general web research. Example intent: "Find the latest docs and announcements about X."
  • Use web_search_advanced_exa for precision research when you must constrain by domains, freshness, depth, or richer query controls. Example intent: "Search only official docs and recent sources for X."
  • Use people_search_exa when the task is about identifying people, roles, background, or profile-style information. Example intent: "Find people working on X at company Y."

Selection Rules

  • Start with web_search_exa when speed and coverage matter more than strict filtering.
  • Choose web_search_advanced_exa when web_search_exa returns noisy or insufficiently targeted results.
  • Choose people_search_exa only for person-centric requests, not generic topic research.

Config Reference

  • Read references/exa-mcp-setup.md for URL templates and generic MCP snippets.
安全使用建议
This appears safe to use if you intend to use Exa MCP for web and people research. Configure the Exa API key carefully, avoid hardcoding or sharing it, and avoid sending private or sensitive search topics to the hosted service unless that matches your privacy requirements.
功能分析
Type: OpenClaw Skill Name: maverick-exa Version: 1.0.0 The skill bundle provides a legitimate integration for the Exa search engine via the Model Context Protocol (MCP). The instructions in SKILL.md and references/exa-mcp-setup.md are well-defined, focusing on web and people research using the official mcp.exa.ai endpoint. It correctly utilizes environment variables (${MAVERICK_EXA_API_KEY}) for secret management and contains no evidence of prompt injection, data exfiltration, or unauthorized execution.
能力标签
requires-sensitive-credentials
能力评估
Purpose & Capability
The stated purpose, SKILL.md workflow, OpenAI prompt, and reference file all align around Exa MCP web search, advanced web search, people search, and setup guidance.
Instruction Scope
The skill limits normal use to three named Exa tools, tells the agent to clarify scope, return source-linked results, and keep other Exa tools disabled unless the user explicitly asks to expand.
Install Mechanism
There is no code or install script. The setup reference provides a hosted MCP URL using an Exa API key placeholder, while registry metadata does not declare required env vars or a primary credential.
Credentials
The skill does not request local file, shell, or system access, but it does route research and people-search requests to the hosted Exa MCP service, which is proportionate to the purpose.
Persistence & Privilege
No persistence, background process, self-propagation, local credential harvesting, or broad account mutation is shown; credential use is limited to the Exa API key placeholder for the hosted endpoint.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install maverick-exa
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /maverick-exa 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.0
Update skill bundle
元数据
Slug maverick-exa
版本 1.0.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 1
常见问题

Maverick Exa 是什么?

Exa MCP workflow for web and people research plus hosted MCP server setup using a fixed tool set (`web_search_exa`, `web_search_advanced_exa`, `people_search... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 60 次。

如何安装 Maverick Exa?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install maverick-exa」即可一键安装,无需额外配置。

Maverick Exa 是免费的吗?

是的,Maverick Exa 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Maverick Exa 支持哪些平台?

Maverick Exa 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 Maverick Exa?

由 Maverick(@maverick)开发并维护,当前版本 v1.0.0。

💬 留言讨论