全网搜索
chanai-search-workflow
by cfyuanyi
External web search workflow for Chinese and global public-web information. Use when searching news, flights, prices, products, docs, tutorials, community posts, websites, PDFs, research, or other external information. Classify intent first, detect dynamic real-time topics, choose domestic-first or global-first search routes, prioritize human-usable sites, generate starter URLs, apply fallback strategy, and report reliability.
安装
claude skill add --url github.com/openclaw/skills/tree/main/skills/cfyuanyi/chanai-search-workflow文档
Chanai Search Workflow
Use this skill as the default workflow for external/public-web search.
Do not use it for local memory recall; use memory search for that.
Core workflow
- Run
scripts/run_search_workflow.py "<query>"first. - Read the returned plan:
- intent
- route (
domestic-first/global-first/mixed) - dynamic subtype if any
- suggested site priority
- starter URLs
- Execute search using the highest-priority routes first.
- If the first route fails, read
references/fallbacks.mdand switch engines/sites. - After collecting evidence, use
scripts/score_result.pyto score reliability. - Use
scripts/search_report.pyorreferences/reporting-template.mdto produce the final response.
Dynamic topics rule
If the query is about flights, prices, inventory, tickets, bookings, hotel availability, or other real-time topics:
- treat public search as preliminary discovery only
- prefer real-time/official/transaction-adjacent pages first
- do not present snippets or aggregator blurbs as final truth
- distinguish clearly between preliminary findings and real-time page findings
Scripts
Unified entrypoint
python3 scripts/run_search_workflow.py "上海直飞赤峰航班价格"
python3 scripts/run_search_workflow.py "openclaw memory search docs"
Individual helpers
python3 scripts/query_router.py "周杰伦演唱会门票开售"
python3 scripts/dynamic_guard.py "上海外滩酒店房价"
python3 scripts/build_urls.py domestic-first dynamic "iPhone 16 京东价格"
python3 scripts/score_result.py 5 5 4 5
python3 scripts/search_report.py domestic-first "Ctrip" "Trip" medium "查到了候选结果" "公开搜索只做初筛" "最终以实时页为准"
References
references/dynamic-info.md— boundary rules for real-time topicsreferences/priorities.md— source priority by scenarioreferences/intent-classification.md— intent taxonomyreferences/direct-templates.md— direct site-entry URL templatesreferences/use-cases.md— reusable search patternsreferences/fallbacks.md— fallback strategy when search is blocked/weakreferences/reliability-scoring.md— reliability rubricreferences/reporting-template.md— final response templatesreferences/domestic-search.md— Chinese-site search guidancereferences/international-search.md— global-site search guidance
Notes
- Prefer domestic-first for Chinese consumer info, local travel, Chinese community content, and local ecommerce.
- Prefer global-first for GitHub, API docs, English tutorials, research, and international sources.
- Prefer mixed when both Chinese and global perspectives matter.
- For dynamic subtypes, the workflow should prioritize scenario-specific sources such as flights, hotels, product-price, ticketing, and finance-news.
相关 Skills
技能工坊
by anthropics
覆盖 Skill 从创建到迭代优化全流程:起草能力、补测试提示、跑评测与基准方差分析,并持续改写内容和描述,提升效果与触发准确率。
✎ 技能工坊把技能从创建、迭代到评测串成闭环,方差分析加描述优化,特别适合把触发准确率打磨得更稳。
PPT处理
by anthropics
处理 .pptx 全流程:创建演示文稿、提取和解析幻灯片内容、批量修改现有文件,支持模板套用、合并拆分、备注评论与版式调整。
✎ 涉及PPTX的创建、解析、修改到合并拆分都能一站搞定,连备注、模板和评论也能处理,做演示文稿特别省心。
PDF处理
by anthropics
遇到 PDF 读写、文本表格提取、合并拆分、旋转加水印、表单填写或加解密时直接用它,也能提取图片、生成新 PDF,并把扫描件通过 OCR 变成可搜索文档。
✎ PDF杂活别再来回切工具了,文本表格提取、合并拆分到OCR识别一次搞定,连扫描件也能变可搜索。
相关 MCP 服务
文件系统
编辑精选by Anthropic
Filesystem 是 MCP 官方参考服务器,让 LLM 安全读写本地文件系统。
✎ 这个服务器解决了让 Claude 直接操作本地文件的痛点,比如自动整理文档或生成代码文件。适合需要自动化文件处理的开发者,但注意它只是参考实现,生产环境需自行加固安全。
by wonderwhy-er
Desktop Commander 是让 AI 直接执行终端命令、管理文件和进程的 MCP 服务器。
✎ 这工具解决了 AI 无法直接操作本地环境的痛点,适合需要自动化脚本调试或文件批量处理的开发者。它能让你用自然语言指挥终端,但权限控制需谨慎,毕竟让 AI 执行 rm -rf 可不是闹着玩的。
EdgarTools
编辑精选by dgunning
EdgarTools 是无需 API 密钥即可解析 SEC EDGAR 财报的开源 Python 库。
✎ 这个工具解决了金融数据获取的痛点——直接让 AI 读取结构化财报,比如让 Claude 分析苹果的 10-K 文件。适合量化分析师或金融开发者快速构建数据管道。但注意,它依赖 SEC 网站稳定性,高峰期可能延迟。