ultimate-search
UltimateSearch
为 Pi/OpenClaw agent 提供双引擎网络搜索能力：Grok AI 搜索（实时联网 + AI 分析）+ Tavily 搜索（结构化结果 + 网页抓取）。
可用工具
在 Bash 中调用以下脚本（无需加入 PATH，脚本会自动读取 .env）：
| 工具 | 命令 | 用途 |
|---|---|---|
| Grok 搜索 | {baseDir}/scripts/grok-search.sh --query "..." |
AI 驱动的深度搜索，Grok 自带联网，返回综合分析 |
| Tavily 搜索 | {baseDir}/scripts/tavily-search.sh --query "..." |
结构化搜索结果，带评分和排序 |
| 网页抓取 | {baseDir}/scripts/web-fetch.sh --url "..." |
提取指定 URL 的完整内容，返回 Markdown |
| 站点映射 | {baseDir}/scripts/web-map.sh --url "..." |
发现网站结构，获取所有 URL |
| 双引擎搜索 | {baseDir}/scripts/dual-search.sh --query "..." |
并行执行 Grok + Tavily，交叉验证 |
各工具参数详见 --help。
More from hamsterider-m/personal-skills
document-parser
文档解析器。支持 PDF/DOCX/EPUB 等格式，返回提取的文本内容。
1deep-learning
Orchestrate deep learning workflows using NotebookLM. Triggers when user wants to deeply understand a topic, research a concept, or learn from materials (links, videos, documents, keywords). Automatically collects materials, uploads to NotebookLM, generates multi-format artifacts (report/podcast/slides/video/quiz/flashcards), and delivers all via Feishu.
1browser-plus
Intelligent browser automation with smart routing between OpenClaw native browser tool and Vercel agent-browser. Use when needing to interact with web pages, especially for filling forms, clicking elements, or typing text in rich text editors that resist standard DOM manipulation. Automatically detects element types and routes to the best execution strategy.
1issue-driven
Issue-driven development workflow. Manage GitHub Issues as SSOT for all dev tasks — create, worktree, code review, PR, and status tracking.
1content-bridge
通用内容摄取路由层。自动识别内容源类型（微信、YouTube、网页、文档等），调用对应的摄取 skill，返回标准化 Markdown 内容。可被任何需要内容摄取的 skill 复用，不绑定特定下游应用。
1web-reader
Extract clean text content from any webpage using Defuddle (primary) or Jina AI (fallback). Use when needing to fetch article content, social media posts, or any web page text without dealing with paywalls, login walls, or complex HTML parsing. Automatically converts web pages to clean Markdown format.
1