patent-reviewer ยท diff

git:20261111.1ff339d to git:20261118.818cff9

4 added, 2 removed. Audit A to A.

What this update changes in the checks

No check changed: every fixed check gives the same answer before and after this update.

A tripped check is a label, not a verdict: read the matched text and the lines below. This week's updates that newly trip a check.

---
name: patent-reviewer
description: Expert system for reviewing utility patent applications against USPTO MPEP guidelines.
---
# Patent Creator Skill
Antecedent checking is disabled by default. Report it as skipped unless `checks_run` includes `antecedent_basis`; retain the reason in `checks_skipped`. Enabling `PATENT_ENABLE_ANTECEDENT_CHECK=1` requires a server restart. Even when enabled, this is a low-confidence pattern check requiring manual review.
Comprehensive patent creation system with USPTO MPEP, prior art databases, and USPTO API for complete patent application analysis.
## Patent search providers
- Use `check_google_patents_status` first, then `search_patents_google` when SerpApi is configured. BigQuery is optional: call any `*_bigquery` tool only after `check_bigquery_status` confirms it is available. Without BigQuery, read the returned patent links for details, CPC codes and family members; use Google Patents keyword/CPC queries for follow-up searches. Mark unavailable details, family coverage and classification searches as unverified. If neither provider is configured, say patent search needs one free key (README Quick Start, Step 3) and run a no-key first screen in the meantime: several `WebSearch` queries restricted to Google Patents (`allowed_domains: ["patents.google.com"]`), then read the closest patent pages with `WebFetch` for claims, dates, CPC codes and citations. Label everything found this way a web first screen, not a prior-art search: web search ranks pages rather than matching claim text and misses much of the record, so a thin or empty result is never evidence of no prior art. A first screen never supports a novelty, obviousness or patentability verdict: in every later assessment, report section or phase that would give one (including any NOVEL / LIKELY NO verdict), write "Not assessed: only a web first screen was run; set up patent search and rerun" instead.
+ Use `check_google_patents_status` first, then `search_patents_google` when SerpApi is configured. If Google Patents is not ready, call `check_epo_api_status`: when it is ready, `search_epo_patents` searches worldwide patent titles and abstracts (CQL `ta=` keywords) and classes (`cpc=` or `ic=` codes), and `get_epo_patent` retrieves claims and description where OPS supplies full text (EP, WO and selected national publications; not every worldwide hit). Check the actual text: empty fields or "not found/not available" messages are not a successful read. Successful OPS keyword/class searches count as patent search, with coverage limits recorded. `search_uspto_api` (after `check_uspto_api_status`) adds US application records via ODP application metadata queries (including titles, parties and status), without claims/description full-text search; use it only as a supplement. BigQuery is optional: call any `*_bigquery` tool only after `check_bigquery_status` confirms it is available. Without BigQuery, use the ready provider: Google Patents links and keyword/CPC queries, or OPS results, `get_epo_patent`, `get_epo_patent_family` and CQL keyword/class queries. OPS search results contain publication numbers, not document URLs. Mark unavailable details, family coverage and classification searches as unverified. If none of Google Patents, EPO OPS and BigQuery is ready, report the status errors and explain how to restore a provider or set up SerpApi (README Quick Start, Step 3) and run a no-key first screen in the meantime: several `WebSearch` queries restricted to Google Patents (`allowed_domains: ["patents.google.com"]`), then read the closest patent pages with `WebFetch` for claims, dates, CPC codes and citations. Label everything found this way a web first screen, not a prior-art search: web search ranks pages rather than matching claim text and misses much of the record, so a thin or empty result is never evidence of no prior art. A first screen never supports a novelty, obviousness or patentability verdict: in every later assessment, report section or phase that would give one (including any NOVEL / LIKELY NO verdict), write "Not assessed: only a web first screen was run; set up patent search and rerun" instead.
## When to Use
Review patent applications for USPTO compliance, analyze claims/specifications/formalities, integrate prior art, get USPTO guidance, assist with patent drafting.
## Quick Review Commands
```bash
/full-review # Complete parallel review
/review-claims # 35 USC 112b compliance
/review-specification # 35 USC 112a compliance
/review-formalities # MPEP 608 compliance
/create-patent # New patent application
```
## Available MCP Tools
### MPEP & Regulations
- `search_mpep` - Search MPEP, 35 USC, 37 CFR
- `get_mpep_section` - Retrieve indexed section content by number. Check `truncated`
and `coverage`; the default caps output at 50 chunks and subsection boundaries
are approximate. Increase `max_chunks` if needed, and do not claim complete text.
### Patent Search
- `search_patents_google` - Full-text search worldwide, claims included (recommended)
+ - `check_epo_api_status`, `search_epo_patents` - OPS fallback: worldwide title/abstract (`ta=`) and class (`cpc=` / `ic=`) search
+ - `get_epo_patent`, `get_epo_patent_family` - Available text and family records
- `search_patents_bigquery` - Fallback keyword search (~$2 a search)
- `get_patent_bigquery` - Get patent details (claims by default; abstract/description opt-in)
- `get_patents_bigquery` - Details for up to 50 patents in one query, same cost as one
- `search_patents_by_cpc_bigquery` - Search by CPC classification
### Patent Analysis
- `review_patent_claims` - Analyze claims for 35 USC 112(b)
- `review_specification` - Check specification support (112a)
- `check_formalities` - Verify MPEP 608 compliance
### Diagram Generation
- `render_diagram` - Create diagrams from DOT code
- `create_flowchart` - Generate patent-style flowcharts
- `create_block_diagram` - Create system block diagrams
- `add_diagram_references` - Add reference numbers
## Review Workflows
### Complete Patent Creation Review (/full-review)
Runs all analyzers in **parallel** for comprehensive analysis:
**Output:**
- All compliance issues across components
- Severity ratings (critical/important/minor)
- Specific MPEP citations
- Actionable fix recommendations
- Prioritized remediation plan
### Claims-Only Review (/review-claims)
**35 USC 112(b) Compliance:**
- Antecedent basis
- Definiteness
- Claim structure
- Subjective terms
- Means-plus-function: separate manual review with the specification required
### Specification Review (/review-specification)
**35 USC 112(a) Requirements:**
- Written description
- Enablement
- Best mode: manual review only; not implemented by the analyzer
- Claim support
### Formalities Check (/review-formalities)
**MPEP 608 Compliance:**
- Abstract (50-150 words)
- Title (<=500 characters)
- Drawing references
- Required sections
## Patent Creation Workflow (/create-patent)
Complete 6-phase patent drafting (55-80 minutes):
1. **Discovery (10-15 min)** - Gather invention details
2. **Technology Analysis (5 min)** - Assess patentability (101, 102, 103)
3. **Specification Drafting (15-20 min)** - Background, summary, detailed description
4. **Claims Drafting (10-15 min)** - Independent + dependent claims
5. **Diagrams & Abstract (10-15 min)** - Block diagrams, flowcharts, abstract
6. **Automatic Validation (5-10 min)** - Runs /full-review, provides fixes
**Output:** USPTO-ready filing package with diagrams
## MPEP Research
```python
# General search
search_mpep("claim definiteness 112(b)", top_k=5)
# Filtered by source
search_mpep("enablement", source_filter="35_USC")
search_mpep("abstract", source_filter="MPEP")
# Get specific section
get_mpep_section("2173") # Claim definiteness
```
### Common MPEP Sections
| Section | Topic |
|---------|-------|
| 608 | Formalities (abstract, title, drawings) |
| 2100 | Patentability requirements |
| 2163 | Guidelines for 35 USC 112(a) |
| 2173 | Claim definiteness (35 USC 112(b)) |
## Prior Art Integration
```python
# Full-text search, worldwide, claims included
- # (fallback: search_patents_bigquery with the same arguments)
+ # If unavailable, follow the OPS CQL fallback above; BigQuery is optional.
search_patents_google(
query="neural network training",
country="US",
start_year=2020,
limit=20
)
# Optional CPC classification search, only after check_bigquery_status succeeds
search_patents_by_cpc_bigquery(cpc_code="G06N3", limit=50)
```
**Integrate findings:**
1. Cite in Background section
2. Emphasize distinctions in Summary
3. Explain advantages in Detailed Description
4. Draft claims to avoid/distinguish
5. List in IDS
## Best Practices
**Before Review:**
- Prepare complete application
- Run /full-review
- Address critical issues first
**During Review:**
- Focus on critical issues (antecedent basis, claim support, definiteness)
- Use MPEP citations
- Iterate until compliant
**After Review:**
- Document compliance
- Final /full-review validation
- Prepare filing package
## Common Review Findings
**Critical (Must Fix):**
- Missing antecedent basis
- Claim elements unsupported
- Abstract exceeds 150 words
- Indefinite language
**Important (Should Fix):**
- Subjective terms without criteria
- Weak enablement
- Inconsistent terminology
**Minor (Optional):**
- Add example embodiments
- Strengthen best mode
- Improve claim scope
## Quick Reference
### Key Compliance Checks
| Requirement | Citation | Tool |
|-------------|----------|------|
| Antecedent basis | 35 USC 112(b) | review_patent_claims |
| Written description | 35 USC 112(a) | review_specification |
| Enablement | 35 USC 112(a) | review_specification |
| Abstract length | MPEP 608.01(b) | check_formalities |
| Title format | MPEP 606 | check_formalities |