professional-patent-agents（专业专利 Agent 人）

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

Boundary: Does not handle Office Action (OA) responses - leave that to professional patent agencies.

Dependencies

Required Skills

Python Dependencies

pip install requests python-docx

Agent List (9 Core Agents)

Workflows

Scenario 1: User Idea → Drafting + Search

flowchart TB
    subgraph S1[Phase 1: Tech Mining]
        M1[tech-miner<br/>Understand idea] ~~~ M2[Extract innovations] ~~~ M3[Disclosure framework]
    end
    
    subgraph S2[Phase 2: Search]
        R1[prior-art-researcher<br/>Keyword strategy] ~~~ R2[Multi-source search] ~~~ R3[Analyze results]
    end
    
    subgraph S3[Phase 3: Evaluation]
        E1[inventiveness-evaluator<br/>Inventiveness analysis] ~~~ E2[Risk scoring] ~~~ E3[Differentiation advice]
    end
    
    subgraph S4[Phase 4: Drafting]
        D1[patent-drafter<br/>Draft specification] ~~~ C1[claims-architect<br/>Design claims]
    end
    
    subgraph S5[Phase 5: Review]
        A1[patent-auditor<br/>Full review] ~~~ A2[Grant rate prediction]
    end
    
    S1 --> S2 --> S3 --> S4 --> S5

Trigger:

"Help me write a patent: [technical idea]"
"I have an idea and want to apply for a patent"

Output Files:

• TECH_DISCLOSURE.md - Technical disclosure framework
• KEYWORD_STRATEGY.md - Keyword strategy
• PATENT_SEARCH_REPORT.md - Search report
• INVENTIVENESS_REPORT.md - Inventiveness evaluation report
• Patent-*.md - Complete patent document (7-section standard format)
• PATENT_AUDIT_REPORT.md - Audit report (with grant rate prediction)
• Patent-*.docx - Word document (auto-converted)

Scenario 2: User Draft → Optimization

flowchart TB
    subgraph P1[Phase 1: Analysis]
        A1[patent-analyst<br/>Parse draft] ~~~ A2[Identify issues] ~~~ A3[Extract keywords]
    end
    
    subgraph P2[Phase 2: Search]
        R1[prior-art-researcher<br/>Targeted search] ~~~ R2[Analyze results]
    end
    
    subgraph P3[Phase 3: Evaluation]
        E1[inventiveness-evaluator<br/>Inventiveness risk] ~~~ E2[Enhancement suggestions]
    end
    
    subgraph P4[Phase 4: Optimization]
        D1[patent-drafter<br/>Optimize specification] ~~~ C1[claims-architect<br/>Strengthen claims]
    end
    
    subgraph P5[Phase 5: Review]
        Q1[patent-auditor<br/>Full review] ~~~ Q2[Grant rate prediction]
    end
    
    P1 --> P2 --> P3 --> P4 --> P5

Trigger:

"Help me optimize this patent: /path/to/patent.md"
"Review this patent and provide optimization suggestions"

Output Files:

• PATENT_ANALYSIS_REPORT.md - Analysis report
• PATENT_SEARCH_REPORT.md - Search report
• INVENTIVENESS_REPORT.md - Inventiveness evaluation report
• PATENT_OPTIMIZATION_SUGGESTIONS.md - Optimization suggestions
• Patent-*.md - Optimized patent document
• PATENT_AUDIT_REPORT.md - 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.

flowchart TB
    subgraph F1[Phase 1: Read Feedback]
        R1[Read search report] ~~~ R2[Read prior art] ~~~ R3[Read original draft]
    end
    
    subgraph F2[Phase 2: Comparative Analysis]
        A1[inventiveness-evaluator<br/>Comparison] ~~~ A2[Identify conflicts] ~~~ A3[Assess differentiation space]
    end
    
    subgraph F3[Phase 3: Decision]
        D1{Inventiveness space?}
        D2[🟢 Sufficient<br/>Recommend optimization]
        D3[🟡 Limited<br/>User confirmation needed]
        D4[🔴 No space<br/>Recommend cancellation]
        D1 --> D2
        D1 --> D3
        D1 --> D4
    end
    
    subgraph F4[Phase 4: Optimization]
        O1[patent-drafter<br/>Targeted optimization] ~~~ O2[Strengthen differences] ~~~ O3[patent-auditor<br/>Final review]
    end
    
    F1 --> F2 --> F3
    D2 --> F4
    D3 -->|User confirms| F4
    D4 --> X[Output cancellation report]

Trigger:

"The agency gave me a search report, help me see if I can optimize: /path/to/report.pdf"
"Here's the prior art, help me evaluate if I need to modify the patent"
"The agency says there's risk, should I continue or cancel?"

Output Files:

• AGENCY_FEEDBACK_ANALYSIS.md - Agency feedback analysis
• DECISION_RECOMMENDATION.md - Decision recommendation (continue/cancel)
• PATENT_OPTIMIZATION_SUGGESTIONS.md - Targeted optimization suggestions

Decision Criteria:

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

Use tech-miner to analyze the technical idea:
[User description]

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

Use prior-art-researcher to search:
Keywords: [keywords]
Technical field: [field]

Search Channels (Default):

⚠️ 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:

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)

Use inventiveness-evaluator to evaluate inventiveness:
Patent document: /path/to/patent.md
Search report: PATENT_SEARCH_REPORT.md

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)

Use patent-drafter to draft patent:
Patent title: [title]
Technical disclosure: TECH_DISCLOSURE.md
Search report: PATENT_SEARCH_REPORT.md
Inventiveness evaluation: INVENTIVENESS_REPORT.md

Claims Architect (Claims Architect)

Identity: Claude, 1000+ claims experience

Trigger: Parallel participation during drafting phase

Output: Claims document

Use claims-architect to design claims:
Patent document: /path/to/patent.md
Core inventive points: [points]

Patent Analyst (Patent Analyst)

Identity: Dr. Zhang, 12 years of patent analysis experience

Trigger: First step in optimization scenario

Output: Analysis report + optimization suggestions

Use patent-analyst to analyze patent:
Patent document: /path/to/patent.md

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)

Use patent-auditor to audit patent:
Patent document: /path/to/patent.md

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)

Use patent-value-appraiser to evaluate patent value:
Patent title: [title]
Or patent number: [number]
Purpose: pledge financing / licensing negotiation / technology transfer / M&A transaction

5 Evaluation Dimensions:

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
2. Extract Mermaid code blocks and render to PNG
3. Use Pandoc to convert section content
4. Fill into Word template at corresponding positions
5. Embed images into document
6. Output to same directory as source file

Dependencies:

Standard Patent Template (7 Sections)

# [Patent Title]

## 1. Related Prior Art and Their Defects or Deficiencies
### 1.1 Description of Prior Art
### 1.2 Defects or Deficiencies of Prior Art

## 2. Technical Improvements to Overcome the Above Defects

## 3. Alternative Solutions for Technical Improvements

## 4. Detailed Embodiments of the Technical Solution
### 4.1 System Architecture
### 4.2 Signal Logic Relationships
### 4.3 Implemented Functions
### 4.4 Specific Implementation Steps
### 4.5 Embodiment 1
(Describe included components/modules and their connections, signal logic; implemented functions and specific implementation steps)
## 5. Advantages of This Proposal Over Prior Art
(Achieved technical effects)
## 6. Related Drawings
(Structure diagrams with labeled component/module names; flowcharts with clear steps and process directions)
## 7. Claims
(Optional)

Language Adaptation

The agents automatically detect and use the user's language for output.

Chinese 7-Section Template (中文七章节模板)

# [专利名称]

## 一、相关的现有技术及现有技术的缺陷或不足
### 1.1 现有技术描述
### 1.2 现有技术的缺陷或不足

## 二、为克服上述缺陷本提案的技术改进点

## 三、技术改进点的其他替代方案

## 四、详细的技术方案具体实施例
### 4.1 系统架构
### 4.2 信号逻辑关系
### 4.3 实现功能
### 4.4 具体实现步骤
### 4.5 实施例一

## 五、本提案相对现有技术的优点

## 六、相关附图

## 七、权利要求书

Key Rules for All Languages

1. 7-Section Structure — Must follow the standard template regardless of language
2. No Executable Code — Use pseudocode or flowcharts instead
3. Quantified Technical Effects — Always quantify improvements (e.g., "30% efficiency increase")
4. Comparison Table — Include comparison with prior art
5. Mermaid Diagrams — Use flowcharts and architecture diagrams

Output Files Summary

Scenario 1: User Idea → Drafting

Scenario 2: User Draft → Optimization

Scenario 3: Agency Feedback → Evaluation

Patent Value Assessment

Installation Location

skills/
└── professional-patent-agents/
    ├── SKILL.md
    ├── agents/
    │   ├── tech-miner/
    │   ├── prior-art-researcher/
    │   ├── inventiveness-evaluator/
    │   ├── patent-drafter/
    │   ├── claims-architect/
    │   ├── patent-analyst/
    │   ├── patent-auditor/
    │   ├── patent-value-appraiser/
    │   └── patent-converter/
    └── skills/
        └── continuous-learning/

Auto-Conversion Flow

Trigger Conditions

Auto-invoke patent-converter to convert Markdown to Word when:

Conversion Flow

flowchart TB
    A[patent-auditor<br/>Review complete] --> B{Grant rate ≥ 60%?}
    B -->|Yes| C[Auto-invoke<br/>patent-converter]
    B -->|No| D[Prompt user confirmation]
    D -->|Confirm convert| C
    D -->|Don't convert| E[Output Markdown only]
    C --> F[Output .docx file]
    F --> G[Notify user<br/>Ready for agency submission]

Output Location

Source directory/
├── Patent-1-xxx.md      # Source Markdown
├── Patent-1-xxx.docx    # Converted output (same directory)
├── Patent-2-xxx.md
├── Patent-2-xxx.docx
└── ...

Innovation Mining & Notifications

Work Record Source

memory/
├── 2026-03-16.md    # Daily work records
├── 2026-03-17.md
├── 2026-03-18.md
└── ...

Innovation Mining Rules

1. Extract technical points from work records: Code commits, architecture designs, problem solutions
2. Combine with industry trends: Search for latest technical developments
3. Evaluate grant rate: Prior art search + Inventiveness evaluation
4. Selection criteria: Grant rate ≥ 65%, Low/Medium risk level

Notification Format

📢 Innovation Mining Report

📊 X potential innovations discovered this week:

1. 【Highly Recommended】A method for XXX
   - Grant rate prediction: 70-80%
   - Risk level: Low
   - Innovation: XXX

2. 【Recommended】A system for XXX
   - Grant rate prediction: 65-75%
   - Risk level: Medium
   - Innovation: XXX

📁 Detailed documents: /patent/weekly/

💡 Suggestion: Prioritize applying for patent #1

⚠️ Security & Installation Guide

System Dependencies

# 1. Install Pandoc (Markdown to Word conversion)
apt install pandoc

# 2. Install Node.js and mermaid-cli (Mermaid diagram rendering)
curl -fsSL https://deb.nodesource.com/setup_18.x | bash -
apt install nodejs
npm install -g @mermaid-js/mermaid-cli

# 3. Install Chromium (optional, for Playwright/Puppeteer)
playwright install chromium

Security Considerations

Recommended Usage

# Option 1: Non-root user (recommended)
python convert_patents.py /path/to/patents

# Option 2: Docker container (isolated environment)
docker run -v /path/to/patents:/data your-image python convert_patents.py /data

Credential Management

This skill depends on other skills that may require API keys or cookies:

Configure credentials securely before use.

*Professional Patent Agents Suite v1.0.2* *Author: BigPiPiHua* *Mail: 775262592@qq.com* *License: MIT* *Updated: 2026-03-31*