界面已本地化;技能介绍与说明文档按所选语言由 AI 自动翻译,代码、命令、路径和专业术语保留原文,以原作者文档为准。
探索技能
设计与创意 · SKILL.md

prior-art-search

prior-art-search

【设计与创意】Search patent databases and academic literature for prior art relevant to an invention. Use when user says "现有技术检索", "prior art search", "专利检索", "check patents", or wants to find relevant prior art.

独立 GitHub 快照 · 自动分类 · 以原作者文档为准

技能说明

GitHub

name: prior-art-search description: "Search patent databases and academic literature for prior art relevant to an invention. Use when user says "现有技术检索", "prior art search", "专利检索", "check patents", or wants to find relevant prior art." argument-hint: "[invention-description-or-path]" allowed-tools: Bash(*), Read, Glob, Grep, WebSearch, WebFetch, Write

Prior Art Search

Search patents and literature for prior art relevant to: $ARGUMENTS

Adapted from /research-lit for patent-specific searching.

Constants

  • MAX_PATENT_RESULTS = 20 — Maximum patent documents to analyze in detail
  • MAX_PAPER_RESULTS = 15 — Maximum academic papers to analyze in detail
  • SEARCH_YEARS = 10 — How many years back to search
  • PATENT_DATABASES = "google-patents, espacenet" — Patent databases to search

Inputs

Read the invention description from:

  1. $ARGUMENTS if it contains technical details
  2. patent/INVENTION_BRIEF.md if it exists
  3. INVENTION_BRIEF.md if it exists at project root

Shared References

Load ../shared-references/prior-art-databases.md for search strategy templates and IPC/CPC classification guidance.

Workflow

Step 1: Extract Search Concepts

From the invention description, identify:

  1. Core inventive concept: The primary technical contribution (1-2 sentences)
  2. Technical problem: What problem it solves
  3. Key technical features: 4-6 specific technical elements that define the invention
  4. IPC/CPC classes: Predict relevant classification codes (e.g., G06N, G06F)

Step 2: Patent Search

For EACH search concept, search via:

Google Patents (via WebSearch):

WebSearch: "site:patents.google.com [keywords]"
WebSearch: "[keywords] patent"
  • Try primary keywords + technical problem keywords
  • Search in English regardless of target jurisdiction
  • For CN inventions, also search Chinese keywords via WebSearch

Espacenet (via WebFetch):

  • WebFetch worldwide.espacenet.com/search results for key queries
  • Search by predicted IPC/CPC classes

Assignee/Inventor Search:

  • If known companies/universities work in this area, search their patent portfolios
  • WebSearch: "[assignee name] patent [technical area]"

For each potentially relevant patent found:

  • WebFetch the patent page to extract: title, abstract, representative claims, filing date, assignee, current status
  • Record IPC/CPC classification codes

Step 3: Academic Literature Search

Search the same concepts in academic databases:

  1. Google Scholar (via WebSearch): WebSearch "[keywords] site:scholar.google.com"
  2. arXiv (via /arxiv if available, or WebSearch): Search for preprints
  3. Semantic Scholar (via /semantic-scholar if API key set, or WebSearch)

For each relevant paper found:

  • Extract title, authors, venue, year, key contribution

Step 4: Classification and Analysis

For each reference found, assess:

  1. Relevance: How closely does it relate to the invention?
  2. Overlap Risk: Does it disclose the same or similar technical solution?
    • HIGH: Anticipates one or more claim elements
    • MEDIUM: Discloses a related but different approach
    • LOW: Same general field, different approach
  3. Relationship: Is it anticipating, relevant, or merely background?

Organize results by IPC/CPC classification to see the technical landscape.

Step 5: Freedom-to-Operate Assessment (Preliminary)

Based on the search results:

  • Identify patents with claims that potentially cover the invention
  • Note any expired patents (public domain)
  • Flag areas where claim scope overlap is significant

Disclaimer: This is a preliminary assessment only. A professional freedom-to-operate analysis by a patent attorney is recommended before filing.

Step 6: Output

Write patent/PRIOR_ART_REPORT.md with:

## Prior Art Search Report

### Invention Summary
[1-2 sentence description of the searched invention]

### Search Strategy
- Keywords used: [...]
- IPC/CPC classes searched: [...]
- Databases searched: Google Patents, Espacenet, Google Scholar, arXiv
- Date range: [year] to present

### Patent References Found

| # | Patent No. | Title | Date | Assignee | IPC/CPC | Key Teaching | Overlap Risk |
|---|-----------|-------|------|----------|---------|-------------|-------------|
| 1 | CN... / US... | [title] | [date] | [assignee] | [codes] | [2-3 sentences] | HIGH/MEDIUM/LOW |

### Non-Patent Literature Found

| # | Reference | Title | Authors/Venue | Year | Key Contribution | Relevance |
|---|-----------|-------|--------------|------|-----------------|-----------|
| 1 | [DOI/link] | [title] | [authors] | [year] | [1-2 sentences] | HIGH/MEDIUM/LOW |

### Prior Art Landscape
[Organized by technical approach or IPC class, not just chronological]

### Freedom-to-Operate Preliminary Assessment
[Which existing patents might block the invention? What is the risk level?]

### Recommendations
- Suggested claim scope adjustments based on prior art
- Areas where novelty appears strongest
- References to watch during prosecution

Key Rules

  • Never fabricate patent numbers or citations. Mark uncertain references with [VERIFY].
  • Search in English AND the target jurisdiction language (Chinese for CN).
  • Patent prior art includes everything published before the priority date, not just patents.
  • Academic papers are valid prior art for both novelty and inventive step.
  • Include expired patents -- they are public domain but still relevant for novelty.
我的收藏