← Back to Skills Marketplace
stevenybw

Chukonu Search Patent

by Bowen Yu · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ Security Clean
34
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install chukonu-search-patent
Description
专利搜索:通过 Chukonu remote MCP server 检索中国(CNIPA)与欧洲(EPO)专利。支持标题/摘要/权利要求/全文布尔查询、IPC 五级分类、申请人/发明人、专利类型、日期与引证数值范围等字段化高级检索,单篇取著录项/权项/全文。只读、OAuth 登录、强制分页。
README (SKILL.md)

chukonu-search-patent

通过 Chukonu 的 remote MCP server 检索专利。本 skill 不调用本地二进制,也不需要 API key —— 它把 OpenClaw 接到托管的 chukonu MCP server(OAuth 登录,只读检索)。

一次性接入(首次使用前)

openclaw mcp add chukonu \
  --url https://search.houdutech.cn/mcp/ \
  --transport streamable-http \
  --auth oauth
openclaw mcp login chukonu      # 弹浏览器完成 OAuth(Google 或 WeChat),token 由 host 保管

接入后会出现 3 个只读 tool:patent_searchfetchusage

何时使用本 skill

  • 检索特定主题/技术的专利(关键词或一段技术描述)
  • 已知申请号取著录项 / 权利要求 / 全文
  • 按 IPC 分类、年份、申请人/发明人、专利类型、引证次数等结构化筛选

不适用于:网页/新闻/学术论文搜索(本服务第一版只做专利)。

工具

1. patent_search —— 唯一检索入口

简单关键词(自由文本快路径,等价标题+摘要+权利要求 tiabc):

patent_search(query="图像识别 OR 深度学习", size=10)

高级检索:叠加下列字段化参数(任意组合,至少一个搜索字段;querytiabc 互斥)。文本字段支持 OR 布尔。

类别 参数
文本(支持 OR) title 标题 · abstract_content 摘要 · claim 权利要求 · title_abstract_content · tiabc(标题+摘要+权项) · full(全文)
IPC 五级 class_ipc(精确, 如 G06T11/60) · class_ipc_main 主分类 · class_ipc_section(G) · class_ipc_class(G06) · class_ipc_subclass(G06F) · class_ipc_group(G06F17)
主体 ap 原始申请人 · first_ap 第一申请人 · inventor 发明人 · first_in 第一发明人
号 & 类型 an 申请号 · pn 公开号 · patent_type(逗号分隔; A=发明申请 B=发明授权 U=实用新型 F=外观设计)
范围 application_date='[YYYYMMDD TO YYYYMMDD]'(* 表开放端) · no_ap/no_in/citation_number_of_times/citation_forward_number_of_times='X''X TO Y'
兜底 extra={...} 透传未在 schema 暴露的字段(province/city/publication_date 等)
分页 size 默认 20、上限 100 · cursor 来自上一次响应的 next_cursor · dataset='cn_abstract'(默认) 或 'epo_docdb'

示例:

patent_search(class_ipc_main="G06F OR G06N", first_ap="华为技术有限公司",
              application_date="[20200101 TO 20231231]", patent_type="A,B", size=10)

单条结果 metadata全著录项(IPC 五级 / 申请人 / 发明人 / 各类日期 / 引证计数 / 公开号·申请号 …), 但不含 claims / 全文——取全文用 fetch。给用户呈现时建议保留:申请号、标题、申请人、申请日、IPC、公开号、score。

2. fetch —— 取单篇文档体

fetch(id="patent:cn_abstract:CN202510299981.2", part="record")
  • id 形如 patent:\x3Cdataset>:\x3Capplication_number>,直接用 patent_search 结果里的 id
  • partrecord(默认,全著录项,无 claims) / claims(权项文本,强制分页,重传 next_cursor 取下一段) / description / fulltext(待全文/PDF 语料,当前返回 null+note)。pdf_url 字段恒在(当前 null)。

3. usage —— 查配额

usage()

返回 {quota_total, used, remaining, reset_at_utc, reset_in_seconds}。长 agent run 里建议先查一次, 临近 QUOTA_EXCEEDED 提前收手。

决策准则

  • 用户给主题词/技术描述 → patent_search(query=…),必要时叠加 patent_type / application_date / class_ipc_main
  • 用户给结构化条件(申请人/发明人/IPC/日期/引证)→ patent_search 对应字段组合
  • 用户给申请号 → 先 fetch(id="patent:\x3Cdataset>:\x3Can>", part="record");要权项再 part="claims"
  • 召回很多时用 next_cursor 翻页,而不是反复换词重搜

错误处理

code 应对
UNAUTHORIZED 重跑 openclaw mcp login chukonu 重新授权
QUOTA_EXCEEDED 已达每日配额;告知用户 retry_after/次日重置,或先 usage() 确认
INVALID_ARGUMENT / INVALID_QUERY_SYNTAX 核对字段名/格式(日期 [YYYYMMDD TO YYYYMMDD]patent_type 代码、querytiabc 互斥)后重试
INVALID_CURSOR cursor 失效或 dataset/scope 不一致;不带 cursor 重新检索
BACKEND_UNAVAILABLE 上游暂时不可用,稍后重试
Usage Guidance
Install only if you intend to give an agent ClawHub maintainer and Convex development guidance. Review commands before approving staff moderation, production migration, email, deploy, or full-access review actions, and use least-privilege credentials.
Capability Tags
requires-oauth-tokenrequires-sensitive-credentials
Capability Assessment
Purpose & Capability
The artifacts support ClawHub maintainer workflows, Convex setup/performance/auth/migration work, code review, and moderation; high-impact abilities such as bans, emails, org/package actions, production migrations, and full-access review are purpose-aligned and disclosed.
Instruction Scope
Instructions are scoped to specific ClawHub, GitHub, and Convex tasks, with explicit target, reason, dry-run, verification, and confirmation requirements for destructive or staff actions.
Install Mechanism
The skill references expected project setup commands such as npm/npx installs and local helper scripts; I found no hidden install path, obfuscated loader, or unrelated automatic package execution.
Credentials
Use of GitHub CLI, ClawHub admin CLI, Convex deployments, auth provider setup, and local repo tooling is proportionate to maintainer and development workflows, but should be used only in the intended trusted environment.
Persistence & Privilege
No hidden persistence was evident. The autoreview helper defaults to a nested full-access Codex review and moderation/migration commands can affect production state, but these are disclosed and gated by user direction or confirmation.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install chukonu-search-patent
  3. After installation, invoke the skill by name or use /chukonu-search-patent
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release with the following features: - Patent search via Chukonu remote MCP server, supporting both China (CNIPA) and Europe (EPO) databases - Boolean, fielded, and full-text search on title, abstract, claims, IPC classes, applicant/inventor, type, date, and citation ranges - Retrieve full bibliographic data and specific sections (claims, description) for individual patents - OAuth login required; usage is read-only and quota controlled - Tools included: patent_search, fetch, and usage, with enforced pagination
Metadata
Slug chukonu-search-patent
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is Chukonu Search Patent?

专利搜索:通过 Chukonu remote MCP server 检索中国(CNIPA)与欧洲(EPO)专利。支持标题/摘要/权利要求/全文布尔查询、IPC 五级分类、申请人/发明人、专利类型、日期与引证数值范围等字段化高级检索,单篇取著录项/权项/全文。只读、OAuth 登录、强制分页。 It is an AI Agent Skill for Claude Code / OpenClaw, with 34 downloads so far.

How do I install Chukonu Search Patent?

Run "/install chukonu-search-patent" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Chukonu Search Patent free?

Yes, Chukonu Search Patent is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Chukonu Search Patent support?

Chukonu Search Patent is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Chukonu Search Patent?

It is built and maintained by Bowen Yu (@stevenybw); the current version is v1.0.0.

💬 Comments