首页 › 龙虾技能列表 › Zai Search — 智能搜索

🔍 Zai Search — 智能搜索

v1.0.0

智能搜索工具，支持多引擎搜索和智能结果排序。

1· 922·5 当前·5 累计
by @bartoszpijet (Bartosz Pijet)·MIT-0
下载技能包
License
MIT-0
最后更新
2026/4/8
安全扫描
VirusTotal
可疑
查看报告
OpenClaw
安全
high confidence
该技能的代码和指令与其所述目的（智能搜索）一致，它使用多个搜索引擎进行交互，请求适当的配置，不请求无关的凭证或意外端点。
评估建议
此技能似乎确实做到了它声称的：智能搜索工具，用于多引擎搜索和智能结果排序。安装或运行之前：1) 确认您的搜索API凭证已妥善配置，并且您对授予搜索访问权限感到满意。2) 仅在您信任的环境中运行，避免搜索敏感内容。3) 技能不请求凭证，但注意不要向脚本传递敏感的搜索关键词，除非您信任整个工具链。4) 如果您想要更高的保证，请在运行前在本地审查包含的脚本。...
详细分析 ▾
✓ 用途与能力
Name/description match the implementation: scripts call Z.AI's web_search API and require ZAI_API_KEY. Required binary (node) is expected and proportional.
ℹ 指令范围
SKILL.md instructs only running the included node scripts. The extract script fetches arbitrary URLs and strips HTML (intended behavior for extraction). Note: arbitrary URL fetching can reach internal/metadata endpoints if run in an environment with broad network access — this is a normal capability for a URL-extraction tool but worth being aware of.
✓ 安装机制
No remote install/downloads or package installs are specified. The skill is instruction+local scripts only (no external archive downloads), minimizing install-time risk.
✓ 凭证需求
Only ZAI_API_KEY (primary credential) is required, which is appropriate for calling the Z.AI Web Search API. The code also optionally checks Z_AI_API_KEY as a fallback — benign but documented.
✓ 持久化与权限
always is false and the skill does not request elevated or cross-skill configuration. It does not persist or modify other skills or agent-wide settings.
安全有层次，运行前请审查代码。

License

MIT-0

可自由使用、修改和再分发，无需署名。

运行时依赖

无特殊依赖

版本

latestv1.0.02026/2/19

初始版本，支持智能搜索

● 可疑

安装命令 点击复制

官方npx clawhub@latest install zai-search
镜像加速npx clawhub@latest install zai-search --registry https://cn.clawhub-mirror.com

技能文档

AI-optimized web search using Z.AI Web Search API. Designed for LLMs – returns structured results with titles, URLs, summaries, site names and favicons.

搜索

node {baseDir}/scripts/search.mjs "query"
node {baseDir}/scripts/search.mjs "query" -n 15
node {baseDir}/scripts/search.mjs "query" --domain sohu.com
node {baseDir}/scripts/search.mjs "query" --recency oneWeek
node {baseDir}/scripts/search.mjs "query" --days 7

Options

  • -n <计数>: 数字 的 results (默认: 10, max: 50)
  • --domain : Limit results 到 specified domain (e.g. sohu.com, www.示例.com)
  • --recency <过滤>: 时间 range – oneDay, oneWeek, oneMonth, oneYear, noLimit (默认)
  • --days : Shorthand 对于 recency (1→oneDay, 7→oneWeek, 30→oneMonth, 365→oneYear)

Extract content 从 URL

node {baseDir}/scripts/extract.mjs "https://example.com/article"

Note: Z.AI does not provide an extract API. This script uses native fetch to retrieve pages and strips HTML for basic text extraction. For richer content, use search results (they include summaries).

Setup

数据来源：ClawHub ↗ · 中文优化：龙虾技能库
OpenClaw 技能定制 / 插件定制 / 私有工作流定制

免费技能或插件可能存在安全风险，如需更匹配、更安全的方案，建议联系付费定制

了解定制服务