A strategic search methodology for patent prior art investigation. The 'prior-art-researcher' agent must use this skill's search strategies, classification code system, and analysis frameworks when searching prior art and determining novelty and inventive step. Used for 'prior art search', 'novelty determination', 'inventive step analysis', etc. Note: Claim drafting or specification writing is outside the scope of this skill.
Search strategies, IPC/CPC classification utilization, and analysis frameworks for systematic prior art investigation.
Invention -> Technical Problem -> Solution Means -> Components
Decomposition Matrix:
| Component | Function | Relationship | Tech Field | Core Keywords |
|-----------|----------|-------------|-----------|--------------|
| A | F1 | A->B | Sensor | sensor, detection |
| B | F2 | B->C | AI | neural, learning |
| C | F3 | C->Output | Control | control, adjust |
Keyword Expansion Rules
Core keyword -> Synonyms -> Broader concepts -> Narrower concepts
Example: "artificial intelligence"
Synonyms: AI, machine learning, ML
Broader: data processing, computer system
Narrower: CNN, RNN, transformer, deep learning
Search Query Structure
(Component A keywords OR synonyms) AND
(Component B keywords OR synonyms) AND
(Function/Effect keywords OR synonyms)
Scope adjustment:
Broad search: A AND B (2 elements)
Medium search: A AND B AND C (3 elements)
Narrow search: A AND B AND C AND D (4+ elements)
| Order | Database | Purpose | Coverage |
|---|---|---|---|
| 1 | KIPRIS (Korean Patents) | Domestic prior art | Korean registrations/publications |
| 2 | Google Patents | Global overview | Worldwide patents |
| 3 | Espacenet (EPO) | Europe and PCT | EP, WO |
| 4 | USPTO (US) | US patents | US |
| 5 | Academic DB (Google Scholar) | Non-patent literature | Papers, theses |
| 6 | Web search | Publicly known technology | Products, blogs, news |
| Technical Field | IPC | Description |
|---|---|---|
| IoT/Sensors | G01D, H04W | Measurement, wireless communication |
| AI/Machine Learning | G06N | Computer system-based models |
| Image Processing | G06T, G06V | Image data processing, recognition |
| Blockchain | H04L 9/00 | Cryptographic arrangements |
| Mobile Apps | G06F 3/04 | User interfaces |
| Batteries | H01M | Electrochemical energy |
| Autonomous Driving | B60W, G05D | Vehicle control, route control |
| Biotech | C12N, A61K | Genetic engineering, pharmaceuticals |
| Claim Element | Prior Art 1 | Prior Art 2 | Prior Art 3 |
|-------------- |------------|------------|------------|
| A: [element] | O Same | O Same | X Absent |
| B: [element] | O Same | X Absent | O Same |
| C: [element] | X Absent | O Same | X Absent |
| D: [element] | X Absent | X Absent | X Absent |
Judgment: If all elements exist in a single prior art -> No novelty
If even one is missing -> Novelty may be recognized
T (Teaching): Does the prior art teach the combination?
S (Suggestion): Is there a suggestion to combine in the description?
M (Motivation): Does a person skilled in the art have motivation to combine?
Judgment flow:
1. Select the closest prior art (primary)
2. Identify differentiating elements
3. Search for secondary prior art containing the differentiating elements
4. Determine whether combination motivation exists
5. Check for unexpected effects from the combination
## Prior Art Search Report
### Search Strategy
- Search query: [Described]
- DB: [Databases used]
- Search results: [Number of results]
### Key Prior Art (Top 5)
#### [Prior Art 1] — [Document No.]
- **Title**: [Title of invention]
- **Filing/Publication Date**: [Date]
- **Key Technology**: [Summary]
- **Related Components**: A(O), B(O), C(X), D(X)
- **Differentiation**: [Difference from the present invention]
### Novelty and Inventive Step Comprehensive Opinion
- **Novelty**: [Recognized/Denied] — Basis: [Explanation]
- **Inventive Step**: [Recognized/Questionable] — Basis: [Explanation]
- **Design-Around Direction**: [Suggestion]
references/search-strategy-detail.md