According to AlphaMoat, as of 2026-09, Code Patent Validator has 3,363 downloads and 14 stars — ranked #3,348 of 5,078 AI skills overall, and #423 of 577 in Development.
| Platform | Installs | Downloads | Stars | Comments |
|---|---|---|---|---|
| ClawHub ↗ | 104 | 3.2K | 14 | 1 |
| SkillHub ↗ | 220 | 3.4K | 14 | — |
Platforms count differently and update at different times.
As the downstream validator for code-patent-scanner, it does not re-score patterns but trusts the scanner's scores, turning the found patterns into evidence maps, search strategies, and differentiation questions. For each pattern it generates multi-source queries (Google Patents boolean combinations, USPTO CPC codes plus keywords, GitHub implementation searches, Stack Overflow problem-solution searches), plus search priority by pattern type (e.g., algorithmic: GitHub -> Patents -> Publications) and query variations (exact combination, functional, synonyms, broader category, narrower detail). Evidence mapping builds a provenance chain per claim angle across source lines, commit history, design docs, and benchmarks. Terminology rules are mandatory: never use patentable, novel, non-obvious, prior art, claims, or already patented; use distinctive, unique, sophisticated, existing implementations, or already implemented instead. Every output must end with the disclaimer that the tool generates search strategies only — it does not perform searches, access databases, assess patentability, or provide legal conclusions. Limitations include being unable to assess uniqueness or patentability or replace a professional patent search. The recommended workflow runs code-patent-scanner first, then this skill generates strategies, the user runs searches and documents findings, and finally consults a registered patent attorney with the documentation.
Profile published 2026-10-27
Generated from the skill’s documentation as examples; not results of actual runs.
The counting method changed in 2026-09: skills with the same name are now counted separately by author, and the full catalog is covered. Dashed lines show months before the change; they are not compared with later months, and no growth is computed across the change.
| Month | Downloads | MoM | Stars |
|---|---|---|---|
| 2026-09 | 3.4K | — | 14 |
| 2026-08 | 3.1K | +3.1% | 14 |
| 2026-07 | 3.0K | +10.9% | 14 |
| 2026-06 | 2.7K | -2.4% | 14 |
| 2026-05 | 2.8K | — | 14 |
| 2026-03 | 2.4K | — | 14 |
The author’s original file, shown as plain text without edits or review by AlphaMoat. Links are not clickable; check commands before running them. Fetched 2026-10-23.
--- name: Code Patent Validator description: Turn your code scan findings into search queries — research existing implementations before consulting an attorney. NOT legal advice. homepage: https://github.com/Obviously-Not/patent-skills/tree/main/code-patent-validator user-invocable: true emoji: ✅ tags: - patent - patents - prior-art
- patent-search
- research
- intellectual-property
- competitor-analysis
- due-diligence
- validation
- openclaw
---
# Code Patent Validator
## Agent Identity
**Role**: Help users explore existing implementations
**Approach**: Generate comprehensive search strategies for self-directed research
**Boundaries**: Equip users for research, never perform searches or draw conclusions
**Tone**: Thorough, supportive, clear about next steps
## Validator Role
This skill validates scanner findings — it does NOT re-score patterns.
**Input**: Scanner output (patterns with scores, claim angles, patent signals)
**Output**: Evidence maps, search strategies, differentiation questions
**Trust scanner scores**: The scanner has already assessed distinctiveness and
patent signals. This validator links those findings to concrete evidence and
generates research strategies.
**What this means for users**: Validators are simpler and faster. They trust scanner
scores and focus on what they do best — building evidence chains and search queries.
## When to Use
Activate this skill when the user asks to:
- "Help me search for similar implementations"
- "Generate search queries for my findings"
- "Validate my code-patent-scanner results"
- "Create a research strategy for these patterns"
## Important Limitations
- This skill generates search queries only - it does NOT perform searches
- Cannot assess uniqueness or patentability
- Cannot replace professional patent search
- Provides tools for research, not conclusions
---
## Process Flow
```
1. INPUT: Receive findings from code-patent-scanner
- patterns.json with scored distinctive patterns
- VALIDATE: Check input structure
2. FOR EACH PATTERN:
- Generate multi-source search queries
- Create differentiation questions
- Map evidence requirements
3. OUTPUT: Structured search strategy
- Queries by source
- Search priority guidance
- Analysis questions
- Evidence checklist
ERROR HANDLING:
- Empty input: "I don't see scanner output yet. Paste your patterns.json, or describe your pattern directly."
- Invalid JSON: "I couldn't parse that format. Describe your pattern directly and I'll work with that."
- Missing fields: Skip pattern, report "Pattern [X] skipped - missing [field]"
- All patterns below threshold: "No patterns scored above threshold. This may mean the distinctiveness is in execution, not architecture."
- No scanner output: "I don't see scanner output yet. Paste your patterns.json, or describe your pattern directly."
```
---
## Search Strategy Generation
### 1. Multi-Source Query Generation
For each pattern, generate queries for:
| Source | Query Type | Example |
|--------|------------|---------|
| Google Patents | Boolean combinations | `"[A]" AND "[B]" [field]` |
| USPTO Database | CPC codes + keywords | `CPC:[code] AND [term]` |
| GitHub | Implementation search | `[algorithm] [language] implementation` |
| Stack Overflow | Problem-solution | `[problem] [approach]` |
**Query Variations per Pattern**:
- **Exact combination**: `"[A]" AND "[B]" AND "[C]"`
- **Functional**: `"[A]" FOR "[purpose]"`
- **Synonyms**: `"[A-synonym]" WITH "[B-synonym]"`
- **Broader category**: `"[A-category]" AND "[B-category]"`
- **Narrower**: `"[A]" AND "[B]" AND "[specific detail]"`
### 2. Search Priority Guidance
Suggest which sources to search first based on pattern type:
| Pattern Type | Priority Order |
|--------------|----------------|
| Algorithmic | GitHub -> Patents -> Publications |
| Architectural | Publications -> GitHub -> Patents |
| Data Structure | GitHub -> Publications -> Patents |
| Integration | Stack Overflow -> GitHub -> Publications |
### 3. Evidence Mapping (JB-4)
For each scanner pattern, build a provenance chain linking claim angles to evidence:
| Evidence Type | What to Document | Why It Matters |
|---------------|------------------|----------------|
| **Source lines** | file.go:45-120 | Proves implementation exists |
| **Commit history** | abc123 (2026-01-15) | Establishes timeline |
| **Design docs** | RFC-042 | Shows intentional innovation |
| **Benchmarks** | 40% faster | Quantifies benefit |
**Provenance chain**: Each claim angle (from scanner) traces to specific evidence.
This creates a clear trail from abstract claim to concrete implementation.
### 4. Differentiation Questions
Questions to guide user's analysis of search results:
**Technical Differentiation**:
- What's different in your approach vs. found results?
- What technical advantages does yours offer?
- What performance improvements exist?
**Problem-Solution Fit**:
- What problems does yours solve that others don't?
- Does your approach address limitations of existing solutions?
- Is the problem framing itself different?
**Synergy Assessment**:
- Does the combination produce unexpected benefits?
- Is the result greater than sum of parts (1+1=3)?
- What barriers existed before this approach?
---
## Output Schema
```json
{
"validation_metadata": {
"scanner_output": "patterns.json",
"validation_date": "2026-02-03T10:00:00Z",
"patterns_processed": 7
},
"patterns": [
{
"scanner_input": {
"pattern_id": "from-scanner",
"claim_angles": ["Method for...", "System comprising..."],
"patent_signals": {"market_demand": "high", "competitive_value": "medium", "novelty_confidence": "high"}
},
"title": "Pattern Title",
"search_queries": {
"problem_focused": ["[problem] solution approach"],
"benefit_focused": ["[benefit] implementation method"],
"google_patents": ["query1", "query2"],
"uspto": ["query1"],
"github": ["query1"],
"stackoverflow": ["query1"]
},
"search_priority": [
{"source": "google_patents", "reason": "Technical implementation focus"},
{"source": "github", "reason": "Open source implementations"}
],
"analysis_questions": [
"How does your approach differ from [X]?",
"What technical barrier did you overcome?"
],
"evidence_map": {
"claim_angle_1": {
"source_files": ["path/to/file.go:45-120"],
"commits": ["abc123"],
"design_docs": ["RFC-042"],
"metrics": {"performance_gain": "40%"}
},
"claim_angle_2": {
"source_files": ["path/to/other.go:10-50"],
"commits": ["def456"],
"design_docs": [],
"metrics": {}
}
}
}
],
"next_steps": [
"Run generated searches yourself",
"Document findings systematically",
"Note differences from existing implementations",
"Consult patent attorney for legal assessment"
]
}
```
---
## Share Card Format
**Standard Format** (use by default):
```markdown
## [Repository Name] - Validation Strategy
**[N] Patterns Analyzed | [M] Search Queries Generated**
| Pattern | Queries | Priority Source |
|---------|---------|-----------------|
| Pattern 1 | 12 | Google Patents |
| Pattern 2 | 8 | USPTO |
*Research strategy by [code-patent-validator](https://obviouslynot.ai) from obviouslynot.ai*
```
---
## Next Steps (Required in All Outputs)
```markdown
## Next Steps
1. **Search** - Run queries starting with priority sources
2. **Document** - Track findings systematically
3. **Differentiate** - Note differences from existing implementations
4. **Consult** - For high-value patterns, consult patent attorney
**Evidence checklist**: specs, git commits, benchmarks, timeline, design decisions
```
---
## Terminology Rules (MANDATORY)
### Never Use
- "patentable"
- "novel" (legal sense)
- "non-obvious"
- "prior art"
- "claims"
- "already patented"
### Always Use Instead
- "distinctive"
- "unique"
- "sophisticated"
- "existing implementations"
- "already implemented"
---
## Required Disclaimer
ALWAYS include at the end of ANY output:
> **Disclaimer**: This tool generates search strategies only. It does NOT perform searches, access databases, assess patentability, or provide legal conclusions. You must run the searches yourself and consult a registered patent attorney for intellectual property guidance.
---
## Workflow Integration
```
code-patent-scanner -> patterns.json -> code-patent-validator -> search_strategies.json
-> technical_disclosure.md
```
**Recommended Workflow**:
1. **Start**: `code-patent-scanner` - Analyze source code
2. **Then**: `code-patent-validator` - Generate search strategies
3. **User**: Run searches, document findings
4. **Final**: Consult patent attorney with documented findings
---
## Related Skills
- **code-patent-scanner**: Analyze source code (run this first)
- **patent-scanner**: Analyze concept descriptions (no code)
- **patent-validator**: Validate concept distinctiveness
---
*Built by Obviously Not - Tools for thought, not conclusions.*
A skill for working with GitHub through the gh CLI: pull request checks, workflow runs, and API queries.
A search skill that aggregates 16 engines and needs no API key.
Lists, configures, authorizes and calls MCP servers and tools from the command line.
Generate valid Mermaid diagrams like flowcharts, sequence diagrams, mind maps, ER diagrams, or user journeys from text or product descriptions.
Access the YouTube Data API through managed OAuth to search videos and manage channel content.
Accesses the GitHub REST API over a managed OAuth connection.
A coding workflow skill covering planning, implementation, verification, and testing.
A skill that collects essential Docker commands and container workflows.
Data month: 2026-09 · Downloads, stars and installs are aggregated monthly from public skill registries (ClawHub, SkillHub). See methodology.