Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Systematic Patent Analysis Campaign — 5 strategies for patent landscape analysis, prior art search, white space identification, competitive intelligence, and claim analysis. Produces structured patent intelligence reports.
.claude/skills/yogsoth-ai-patent-mining/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-18 | ✗→✓ | ▲ Improved | -19% | 0% |
| case-19 | ✗→✓ | ▲ Improved | -30% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 20% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 5% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 25% | 0% |
Systematic patent intelligence campaign. Analyzes patent landscapes, traces prior art, identifies white spaces, profiles competitor portfolios, and decomposes claim scope.
| Signal | Strategy | |--------|----------| | "map the patent landscape for X" / "who are the key players in X" / "filing trends" | landscape-survey | | "is X novel" / "find prior art for X" / "freedom to operate" | prior-art-search | | "where are the gaps" / "uncovered areas" / "opportunity mapping" | white-space-analysis | | "what is competitor Y doing" / "portfolio comparison" / "IP strategy" | competitive-intelligence | | "analyze these claims" / "claim scope" / "protection breadth" | claim-analysis |
| Strategy | Budget Summary | |----------|---------------| | landscape-survey | 200 families, 0 claim parses, 80 web searches | | prior-art-search | 80 families, 20 claim parses, 50 web searches | | white-space-analysis | 150 families, 10 claim parses, 60 web searches | | competitive-intelligence | 120 families, 15 claim parses, 40 web searches | | claim-analysis | 30 families, 30 claim parses, 20 web searches |
| Tactic | Purpose | |--------|---------| | patent-family-tracing | Forward/backward citation and priority chain tracing | | classification-navigation | IPC/CPC hierarchy drill-down and lateral expansion | | claim-decomposition | Independent/dependent claim parsing and feature mapping |
| SOP | Purpose | |-----|---------| | patent-query-formulation | Construct multi-faceted patent search strategies | | patent-categorization | Classify patents by subdomain and value chain | | assignee-normalization | Standardize assignee names and corporate groups | | citation-network-analysis | Build citation graphs, compute main path and PageRank | | trend-analysis | Filing volume time-series and S-curve lifecycle | | white-space-mapping | Feature cross-matrix and blank area identification | | claim-parsing | Claim syntax parsing and element extraction | | legal-status-assessment | Determine patent legal status (active/expired/pending) | | quality-scoring | Multi-dimensional patent quality assessment | | patent-synthesis | Produce final structured intelligence report | | saturation-detection | (shared) Determine when search expansion has saturated |
| Strategy | Patent Families | Claim Parses | Web Searches | |----------|----------------|--------------|--------------| | landscape-survey | 200 | 0 | 80 | | prior-art-search | 80 | 20 | 50 | | white-space-analysis | 150 | 10 | 60 | | competitive-intelligence | 120 | 15 | 40 | | claim-analysis | 30 | 30 | 20 |
| Server | Tools Used | Purpose | |--------|-----------|---------| | brave-search | brave_web_search, brave_llm_context | Patent database web search, patent office queries | | apify | google_scholar_scraper | Academic patent literature discovery | | alphaxiv | full_text_papers_search, get_paper_content | Patent-related academic papers | | semantic-scholar | ss_relevance_search, ss_citations, ss_references | Citation network for patent-adjacent literature |
saturation-detection SOP to determine when citation/classification expansion has convergedpatent-synthesis SOP<!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| Strategy | When to use | | --- | --- | | claim-analysis | Deep claim scope analysis — decompose independent/dependent claims and assess protection scope breadth. Budget: 30 patent families, 30 claim parses, 20 web searches. | | competitive-intelligence | Analyze competitor IP portfolios — comparative patent portfolio reports with strategy inference. Budget: 120 patent families, 15 claim parses, 40 web searches. | | landscape-survey | Patent landscape full-scan — maps technology domain via assignee ranking, IPC/CPC classification, filing trends. Budget: 200 patent families, 0 claim parses, 80 web searches. | | prior-art-search | Evaluate novelty of specific invention — find relevant prior art across patents, publications, and products. Budget: 80 patent families, 20 claim parses, 50 web searches. | | white-space-analysis | Identify patent coverage gaps — feature cross-matrix blank areas revealing unprotected technology combinations. Budget: 150 patent families, 10 claim parses, 60 web searches. |
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | context-checkpoint | Append research process and results to the current Phase's context file. Covers both process and results with genuine substance. Use this skill at plan-designated checkpoint points — typically after each strategy completes or at key decision nodes within a research Phase. | | context-init | Create a new context file for a research Phase. Called once at Phase start to initialize the file that subsequent context-checkpoint calls will append to. Use this skill whenever a new research Phase begins and a fresh context file is needed. |
<!-- END available-tables (generated) -->
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-18 | fail→pass | 22,054 | 7,703 | -65% | 1 | 1 | 0% | 2,246 | 1,824 | -19% | 0 | 0 | — |
case-19 | fail→pass | 22,337 | 7,846 | -65% | 1 | 1 | 0% | 2,643 | 1,843 | -30% | 0 | 0 | — |
case-13 | pass→fail | 14,469 | 8,213 | -43% | 1 | 1 | 0% | 1,578 | 1,836 | +16% | 0 | 0 | — |
case-01 | fail→fail | 38,229 | 45,831 | +20% | 1 | 1 | 0% | 5,549 | 2,037 | -63% | 0 | 0 | — |
case-02 | fail→fail | 55,549 | 17,959 | -68% | 1 | 1 | 0% | 7,893 | 1,800 | -77% | 0 | 0 | — |
case-03 | pass→fail | 48,264 | 53,421 | +11% | 1 | 1 | 0% | 7,333 | 1,948 | -73% | 0 | 0 | — |
case-04 | fail→pass | 29,369 | 36,077 | +23% | 1 | 1 | 0% | 3,135 | 3,777 | +20% | 0 | 0 | — |
case-05 | fail→fail | 48,943 | 23,597 | -52% | 1 | 1 | 0% | 3,269 | 2,044 | -37% | 0 | 0 | — |
case-06 | fail→pass | 39,019 | 33,787 | -13% | 1 | 1 | 0% | 2,264 | 2,377 | +5% | 0 | 0 | — |
case-07 | fail→fail | 20,609 | 21,560 | +5% | 1 | 1 | 0% | 2,142 | 1,874 | -13% | 0 | 0 | — |
case-08 | fail→fail | 21,922 | 4,918 | -78% | 1 | 1 | 0% | 2,865 | 2,164 | -24% | 0 | 0 | — |
case-09 | pass→pass | 13,321 | 3,496 | -74% | 1 | 1 | 0% | 1,104 | 1,920 | +74% | 0 | 0 | — |
case-10 | fail→pass | 20,695 | 22,155 | +7% | 1 | 1 | 0% | 2,985 | 3,725 | +25% | 0 | 0 | — |
case-11 | pass→pass | 16,653 | 3,505 | -79% | 1 | 1 | 0% | 2,841 | 1,888 | -34% | 0 | 0 | — |
case-12 | pass→fail | 19,745 | 3,187 | -84% | 1 | 1 | 0% | 2,093 | 1,770 | -15% | 0 | 0 | — |
case-14 | pass→pass | 18,747 | 10,810 | -42% | 1 | 1 | 0% | 2,839 | 3,055 | +8% | 0 | 0 | — |
case-15 | pass→fail | 20,082 | 13,200 | -34% | 1 | 1 | 0% | 2,431 | 2,690 | +11% | 0 | 0 | — |
case-16 | fail→pass | 17,371 | 9,849 | -43% | 1 | 1 | 0% | 2,160 | 2,137 | -1% | 0 | 0 | — |
case-17 | fail→fail | 16,222 | 4,425 | -73% | 1 | 1 | 0% | 1,664 | 1,918 | +15% | 0 | 0 | — |
case-20 | fail→fail | 14,292 | 21,737 | +52% | 1 | 1 | 0% | 2,245 | 1,997 | -11% | 0 | 0 | — |
case-21 | fail→fail | 45,025 | 46,301 | +3% | 1 | 1 | 0% | 7,844 | 9,572 | +22% | 0 | 0 | — |
case-22 | fail→fail | 20,707 | 12,397 | -40% | 1 | 1 | 0% | 3,241 | 1,937 | -40% | 0 | 0 | — |
case-23 | fail→fail | 20,268 | 7,956 | -61% | 1 | 1 | 0% | 3,210 | 1,806 | -44% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted, and 17 counted toward the lift figure. The other 6 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +9 percentage points is the difference between those two pass rates over the 17 comparable cases. 7 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.