Anysearch

MUST USE when the user needs real-time information from the open web: general web search, 18 vertical domains (finance / academic / legal / security / code / travel / health / social_media …), parallel batch search, and full-page extraction. Triggers on 查一下 / 搜一下 / 调研 / 找资料 / 最新消息 / 查行情 / 查财报 / 查论文 / 查专利 / fact-check, and on any request whose answer depends on current data. For structured identifiers (Stock: / CVE: / DOI: / IATA: / patent) discover the sub_domain first and search the vertical domain instead of the general web. Public social-media discovery (X / Reddit / LinkedIn / 微博 / 知乎 / 微信) also runs here, no login needed. NOT for: platform-native / account-scoped data (exact posts, replies, followers, 小红书 / B站 / V2EX / YouTube 字幕) — use agent-reach for platform access; NOT for writing / analysis / translation of content you already have.

xhqing Updated 1 repo stars

File contents

xhqing/agents-infra commit 0d71510f26

Frequently asked questions

npx get-skill add xhqing/anysearch-2