Professional Patent Agents Suite
License
MIT License
Copyright (c) 2026 BigPiPiHua
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
Core Positioning
| Scenario | Input | Output | Goal |
|---|
| Scenario 1 | User idea (vague description) | Complete patent document + Search report | Grant rate + Inventiveness |
| Scenario 2 |
User draft (existing document) | Optimized patent document | Grant rate + Inventiveness |
|
Scenario 3 | Agency feedback (search report/prior art) | Optimization suggestions / Cancellation advice | Decision support |
Boundary: Does not handle Office Action (OA) responses - leave that to professional patent agencies.
Dependencies
Required Skills
| Skill | Purpose | Install Command |
|---|
| INLINECODE0 | AI-optimized search | INLINECODE1 |
| INLINECODE2 |
Academic paper search |
clawhub install aminer-open-academic |
Python Dependencies
CODEBLOCK0
Agent List (9 Core Agents)
| Agent | Role | Core Capability | Priority |
|---|
| tech-miner | Technology Mining Expert | Idea analysis, innovation extraction, technical disclosure framework | ⭐⭐⭐⭐⭐ |
| prior-art-researcher |
Prior Art Search Expert | Keyword strategy + Multi-source search + Analysis | ⭐⭐⭐⭐⭐ |
|
inventiveness-evaluator | Inventiveness Evaluation Expert | Inventiveness analysis, risk scoring, grant rate prediction | ⭐⭐⭐⭐⭐ |
|
patent-drafter | Patent Drafting Expert | 7-section drafting, Mermaid diagrams, document conversion | ⭐⭐⭐⭐⭐ |
|
claims-architect | Claims Architect | Claims design, scope optimization | ⭐⭐⭐⭐ |
|
patent-analyst | Patent Analyst | Draft analysis, issue identification, optimization suggestions | ⭐⭐⭐⭐ |
|
patent-auditor | Patent Audit Expert | Quality review, grant rate prediction, revision suggestions | ⭐⭐⭐⭐⭐ |
|
patent-value-appraiser | Patent Value Appraiser | 5-dimension value assessment, market value estimation | ⭐⭐⭐⭐ |
|
patent-converter | Document Conversion Expert | Markdown→Word, Mermaid diagram embedding | ⭐⭐⭐⭐ |
Workflows
Scenario 1: User Idea → Drafting + Search
CODEBLOCK1
Trigger:
CODEBLOCK2
Output Files:
- -
TECH_DISCLOSURE.md - Technical disclosure framework - INLINECODE5 - Keyword strategy
- INLINECODE6 - Search report
- INLINECODE7 - Inventiveness evaluation report
- INLINECODE8 - Complete patent document (7-section standard format)
- INLINECODE9 - Audit report (with grant rate prediction)
- INLINECODE10 - Word document (auto-converted)
Scenario 2: User Draft → Optimization
CODEBLOCK3
Trigger:
CODEBLOCK4
Output Files:
- -
PATENT_ANALYSIS_REPORT.md - Analysis report - INLINECODE12 - Search report
- INLINECODE13 - Inventiveness evaluation report
- INLINECODE14 - Optimization suggestions
- INLINECODE15 - Optimized patent document
- INLINECODE16 - Audit report
Scenario 3: Agency Feedback → Evaluation
Use Case: User has submitted to a patent agency, received a search report or prior art, needs to evaluate whether to continue optimizing or cancel the application.
CODEBLOCK5
Trigger:
CODEBLOCK6
Output Files:
- -
AGENCY_FEEDBACK_ANALYSIS.md - Agency feedback analysis - INLINECODE18 - Decision recommendation (continue/cancel)
- INLINECODE19 - Targeted optimization suggestions
Decision Criteria:
| Inventiveness Space | Basis | Recommendation |
|---|
| 🟢 Sufficient | Core features not disclosed, clear differentiation points | Continue optimization, strengthen differences |
| 🟡 Limited |
Some features disclosed, need repositioning | User confirmation required |
| 🔴 No space | Core features already disclosed, cannot circumvent | Recommend cancellation or redesign |
Agent Details
Tech Miner (Technology Mining Expert)
Identity: Dr. Li, 15 years of technology assessment and innovation mining experience
Trigger: When user provides vague idea
Output: Technical disclosure framework
CODEBLOCK7
Prior Art Researcher (Prior Art Search Expert)
Identity: Dr. Chen, 15 years of patent search experience
Trigger: When search is needed
Output: Search report + analysis
CODEBLOCK8
Search Channels (Default):
| Priority | Channel | Tool | Purpose |
|---|
| 1 | Tavily | INLINECODE20 | Quick search, technical overview |
| 2 |
AMiner |
aminer-open-academic | Academic papers + patent database |
| 3 | Google Patents |
web_fetch | Global patent full text |
| 4 | GitHub |
tavily site: | Open source projects |
| 5 | Tech blogs |
tavily site: | Technical articles |
⚠️ Patent Database APIs Recommended:
Default channels may not be sufficient for accurate patent prior art search. For professional patent search, recommend users to connect patent database APIs:
| Database | API Type | Coverage | Best For |
|---|
| Google Patents | Public API | 100+ offices | Global search |
| USPTO |
Public API | US patents | US details |
|
EPO (Espacenet) | Public API | European patents | EP search |
|
CNIPA | Public API | Chinese patents | CN search |
|
WIPO | Public API | PCT applications | International |
|
Lens.org | Free API | Global patents | Academic research |
ClawHub Skill Discovery:
# Always check ClawHub for patent search skills
clawhub search patent
clawhub search "prior art"
Inventiveness Evaluator (Inventiveness Evaluation Expert)
Identity: Dr. Zhao, former examiner, 12 years of evaluation experience
Trigger: After search completion
Output: Inventiveness evaluation report (with risk score)
CODEBLOCK10
Patent Drafter (Patent Drafting Expert)
Identity: Patricia, 12 years of drafting experience, 92% grant rate
Trigger: After inventiveness evaluation passes
Output: Complete patent document (7 sections)
CODEBLOCK11
Claims Architect (Claims Architect)
Identity: Claude, 1000+ claims experience
Trigger: Parallel participation during drafting phase
Output: Claims document
CODEBLOCK12
Patent Analyst (Patent Analyst)
Identity: Dr. Zhang, 12 years of patent analysis experience
Trigger: First step in optimization scenario
Output: Analysis report + optimization suggestions
CODEBLOCK13
Patent Auditor (Patent Audit Expert)
Identity: Judge Wu, former examiner, reviewed 3000+ applications
Trigger: After drafting/optimization completion
Output: Audit report (with grant rate prediction)
CODEBLOCK14
Patent Value Appraiser (Patent Value Appraiser)
Identity: Ms. Lin, 10 years of patent valuation experience, certified IP asset appraiser
Trigger: User needs to evaluate existing patent value (transfer, pledge, financing)
Output: Patent value assessment report (5-dimension radar chart + market value range)
CODEBLOCK15
5 Evaluation Dimensions:
| Dimension | Weight | Content |
|---|
| Technical Value | 25% | Innovation degree, technical complexity, substitution difficulty |
| Legal Value |
25% | Claim breadth, stability, invalidation resistance |
| Market Value | 25% | Application scenarios, market size, competitive alternatives |
| Economic Value | 15% | Cost savings, revenue potential, licensing income |
| Strategic Value | 10% | Supply chain position, barrier strength, negotiation leverage |
Patent Converter (Document Conversion Expert)
Identity: Alex, document conversion expert, proficient in Markdown → Word conversion
Trigger: Auto-triggered after patent-auditor review passes
Output: Word document (.docx), with embedded Mermaid diagrams
Conversion Flow:
- 1. Parse 7 sections of patent Markdown
- Extract Mermaid code blocks and render to PNG
- Use Pandoc to convert section content
- Fill into Word template at corresponding positions
- Embed images into document
- Output to same directory as source file
Dependencies:
| Tool | Purpose |
|---|
| Pandoc | Markdown → Word conversion |
| mmdc (mermaid-cli) |
Mermaid → PNG rendering |
| python-docx | Word document operations |
Standard Patent Template (7 Sections)
CODEBLOCK16
Language Adaptation
The agents automatically detect and use the user's language for output.
| User Input Language | Output Language | Template Format |
|---|
| English | English | 7-Section Standard (English) |
| 中文 |
中文 | 7章节标准模板（中文） |
| Other languages | User's language | 7-Section Standard (translated) |
Chinese 7-Section Template (中文七章节模板)
CODEBLOCK17
Key Rules for All Languages
- 1. 7-Section Structure — Must follow the standard template regardless of language
- No Executable Code — Use pseudocode or flowcharts instead
- Quantified Technical Effects — Always quantify improvements (e.g., "30% efficiency increase")
- Comparison Table — Include comparison with prior art
- Mermaid Diagrams — Use flowcharts and architecture diagrams
Output Files Summary
Scenario 1: User Idea → Drafting
| File | Content | Phase |
|---|
| INLINECODE25 | Technical disclosure framework | Tech mining |
| INLINECODE26 |
Keyword strategy | Search |
|
PATENT_SEARCH_REPORT.md | Search report | Search |
|
INVENTIVENESS_REPORT.md | Inventiveness evaluation report | Evaluation |
|
Patent-*.md | Complete patent document | Drafting |
|
PATENT_AUDIT_REPORT.md | Audit report (with grant rate) | Review |
Scenario 2: User Draft → Optimization
| File | Content | Phase |
|---|
| INLINECODE31 | Analysis report | Analysis |
| INLINECODE32 |
Search report | Search |
|
INVENTIVENESS_REPORT.md | Inventiveness evaluation report | Evaluation |
|
PATENT_OPTIMIZATION_SUGGESTIONS.md | Optimization suggestions | Optimization |
|
PATENT_AUDIT_REPORT.md | Audit report (with grant rate) | Review |
Scenario 3: Agency Feedback → Evaluation
| File | Content | Phase |
|---|
| INLINECODE36 | Feedback analysis | Analysis |
| INLINECODE37 |
Decision recommendation | Decision |
|
PATENT_OPTIMIZATION_SUGGESTIONS.md | Optimization suggestions (if chosen) | Optimization |
Patent Value Assessment
| File | Content |
|---|
| INLINECODE39 | 5-dimension value assessment + market value range |
Installation Location
CODEBLOCK18
Auto-Conversion Flow
Trigger Conditions
Auto-invoke patent-converter to convert Markdown to Word when:
| Condition | Requirement |
|---|
| patent-auditor review result | Passed |
| Grant rate prediction |
≥ 60% (low/medium risk) |
| User confirmation | High risk (< 60%) requires user confirmation |
Conversion Flow
CODEBLOCK19
Output Location
CODEBLOCK20
Innovation Mining & Notifications
Work Record Source
CODEBLOCK21
Innovation Mining Rules
- 1. Extract technical points from work records: Code commits, architecture designs, problem solutions
- Combine with industry trends: Search for latest technical developments
- Evaluate grant rate: Prior art search + Inventiveness evaluation
- Selection criteria: Grant rate ≥ 65%, Low/Medium risk level
Notification Format
CODEBLOCK22
⚠️ Security & Installation Guide
System Dependencies
CODEBLOCK23
Security Considerations
| Risk | Description | Recommendation |
|---|
| Puppeteer --no-sandbox | Required when running as root, reduces browser sandboxing | Run in Docker container or non-root user |
| Network requests |
Downloads PDFs and accesses third-party sites | Run in isolated environment |
| Default paths | Converter uses
/root/workspace/patent/new as default | Verify working directory before running |
Recommended Usage
CODEBLOCK24
Credential Management
This skill depends on other skills that may require API keys or cookies:
| Skill | Credentials Needed |
|---|
| tavily-search | Tavily API key |
| aminer-open-academic |
AMiner API key (optional) |
Configure credentials securely before use.
Professional Patent Agents Suite v1.0.2
Author: BigPiPiHua
Mail: 775262592@qq.com
License: MIT
Updated: 2026-03-31专业专利代理人套件
许可证
MIT许可证
版权所有 (c) 2026 BigPiPiHua
特此授权，任何获得本软件及相关文档文件（以下简称软件）副本的人，均可免费使用本软件，不受限制，包括但不限于使用、复制、修改、合并、发布、分发、再许可和/或出售软件副本的权利，并允许获得软件的人员在遵守以下条件的情况下这样做：
上述版权声明和本许可声明应包含在软件的所有副本或实质性部分中。
本软件按原样提供，不提供任何明示或暗示的担保，包括但不限于适销性、特定用途适用性和非侵权性的担保。在任何情况下，作者或版权持有人均不对任何索赔、损害或其他责任负责，无论是在合同、侵权或其他方面，由软件或软件的使用或其他交易引起、由软件引起或与之相关。
核心定位
| 场景 | 输入 | 输出 | 目标 |
|---|
| 场景1 | 用户创意（模糊描述） | 完整专利文档 + 检索报告 | 授权率 + 创造性 |
| 场景2 |
用户草稿（已有文档） | 优化后的专利文档 | 授权率 + 创造性 |
|
场景3 | 代理机构反馈（检索报告/现有技术） | 优化建议 / 撤销建议 | 决策支持 |
边界：不处理审查意见通知书（OA）答复——交由专业专利代理机构处理。
依赖项
必需技能
| 技能 | 用途 | 安装命令 |
|---|
| tavily-search | AI优化搜索 | clawhub install tavily-search |
| aminer-open-academic |
学术论文搜索 | clawhub install aminer-open-academic |
Python依赖项
bash
pip install requests python-docx
智能体列表（9个核心智能体）
| 智能体 | 角色 | 核心能力 | 优先级 |
|---|
| tech-miner | 技术挖掘专家 | 创意分析、创新点提取、技术交底框架 | ⭐⭐⭐⭐⭐ |
| prior-art-researcher |
现有技术检索专家 | 关键词策略 + 多源检索 + 分析 | ⭐⭐⭐⭐⭐ |
|
inventiveness-evaluator | 创造性评估专家 | 创造性分析、风险评分、授权率预测 | ⭐⭐⭐⭐⭐ |
|
patent-drafter | 专利撰写专家 | 七章节撰写、Mermaid图表、文档转换 | ⭐⭐⭐⭐⭐ |
|
claims-architect | 权利要求架构师 | 权利要求设计、范围优化 | ⭐⭐⭐⭐ |
|
patent-analyst | 专利分析师 | 草稿分析、问题识别、优化建议 | ⭐⭐⭐⭐ |
|
patent-auditor | 专利审核专家 | 质量审查、授权率预测、修改建议 | ⭐⭐⭐⭐⭐ |
|
patent-value-appraiser | 专利价值评估师 | 五维度价值评估、市场价值估算 | ⭐⭐⭐⭐ |
|
patent-converter | 文档转换专家 | Markdown→Word、Mermaid图表嵌入 | ⭐⭐⭐⭐ |
工作流程
场景1：用户创意 → 撰写 + 检索
mermaid
flowchart TB
subgraph S1[阶段1：技术挖掘]
M1[tech-miner
理解创意] ~~~ M2[提取创新点] ~~~ M3[交底框架]
end
subgraph S2[阶段2：检索]
R1[prior-art-researcher
关键词策略] ~~~ R2[多源检索] ~~~ R3[分析结果]
end
subgraph S3[阶段3：评估]
E1[inventiveness-evaluator
创造性分析] ~~~ E2[风险评分] ~~~ E3[差异化建议]
end
subgraph S4[阶段4：撰写]
D1[patent-drafter
撰写说明书] ~~~ C1[claims-architect
设计权利要求]
end
subgraph S5[阶段5：审核]
A1[patent-auditor
全面审核] ~~~ A2[授权率预测]
end
S1 --> S2 --> S3 --> S4 --> S5
触发条件：
帮我写一份专利：[技术创意]
我有一个想法，想申请专利
输出文件：
- - TECHDISCLOSURE.md - 技术交底框架
- KEYWORDSTRATEGY.md - 关键词策略
- PATENTSEARCHREPORT.md - 检索报告
- INVENTIVENESSREPORT.md - 创造性评估报告
- Patent-.md - 完整专利文档（七章节标准格式）
- PATENTAUDIT_REPORT.md - 审核报告（含授权率预测）
- Patent-.docx - Word文档（自动转换）
场景2：用户草稿 → 优化
mermaid
flowchart TB
subgraph P1[阶段1：分析]
A1[patent-analyst
解析草稿] ~~~ A2[识别问题] ~~~ A3[提取关键词]
end
subgraph P2[阶段2：检索]
R1[prior-art-researcher
针对性检索] ~~~ R2[分析结果]
end
subgraph P3[阶段3：评估]
E1[inventiveness-evaluator
创造性风险] ~~~ E2[增强建议]
end
subgraph P4[阶段4：优化]
D1[patent-drafter
优化说明书] ~~~ C1[claims-architect
加强权利要求]
end
subgraph P5[阶段5：审核]
Q1[patent-auditor
全面审核] ~~~ Q2[授权率预测]
end
P1 --> P2 --> P3 --> P4 --> P5
触发条件：
帮我优化这份专利：/path/to/patent.md
审核这份专利并提供优化建议
输出文件：
- - PATENTANALYSISREPORT.md - 分析报告
- PATENTSEARCHREPORT.md - 检索报告
- INVENTIVENESSREPORT.md - 创造性评估报告
- PATENTOPTIMIZATIONSUGGESTIONS.md - 优化建议
- Patent-*.md - 优化后的专利文档
- PATENTAUDIT_REPORT.md - 审核报告
场景3：代理机构反馈 → 评估
使用场景：用户已提交至专利代理机构，收到检索报告或现有技术，需要评估是继续优化还是撤销申请。
mermaid
flowchart TB
subgraph F1[阶段1：阅读反馈]
R1[阅读检索报告] ~~~ R2[阅读现有技术] ~~~ R3[阅读原始草稿]
end
subgraph F2[阶段2：对比分析]
A1[inventiveness-evaluator
对比] ~~~ A2[识别冲突] ~~~ A3[评估差异化空间]
end
subgraph F3[阶段3：决策]
D1{创造性空间？}
D2[🟢 充足
建议优化]
D3[🟡 有限
需用户确认]
D4[🔴 无空间
建议撤销]
D1 --> D2
D1 --> D3
D1 --> D4
end
subgraph F4[阶段4：优化]
O1[patent-drafter
针对性优化] ~~~ O2[加强差异化] ~~~ O3[patent-auditor
最终审核]
end
F1 --> F2 --> F3
D2 --> F4
D3 -->|用户确认| F4
D4 --> X[输出撤销报告]
触发条件：
代理机构给了我一份检索报告，帮我看看能否优化：/path/to/report.pdf
这是现有技术，帮我评估是否需要修改专利
代理机构说存在风险，我应该继续还是撤销？
输出文件：
- - AGENCYFEEDBACKANALYSIS.md - 代理机构反馈分析
- DECISIONRECOMMENDATION.md - 决策建议（继续/撤销）
- PATENTOPTIMIZATION_SUGGESTIONS.md - 针对性优化建议
决策标准：
| 创造性空间 | 依据 | 建议 |
|---|
| 🟢 充足 | 核心特征未被公开，差异化点明确 | 继续优化，加强差异化 |
| 🟡 有限 |
部分特征已被公开，需重新定位 | 需用户确认 |
| 🔴 无空间
