assess-ip-landscape
À propos
Cette compétence analyse le paysage de la propriété intellectuelle pour une technologie donnée, en réalisant une analyse des clusters de brevets, une évaluation des concurrents et un examen préliminaire de la liberté d'opération. Elle est conçue pour être utilisée avant de démarrer la R&D, lors de l'évaluation d'une entrée sur le marché ou en préparation d'une diligence raisonnable pour un investissement. L'outil aide à identifier les espaces libres et fournit des recommandations stratégiques de positionnement en propriété intellectuelle.
Installation rapide
Claude Code
Recommandénpx skills add pjt222/agent-almanac -a claude-code/plugin add https://github.com/pjt222/agent-almanacgit clone https://github.com/pjt222/agent-almanac.git ~/.claude/skills/assess-ip-landscapeCopiez et collez cette commande dans Claude Code pour installer cette compétence
Documentation
Assess IP Landscape
Map IP landscape for tech domain — ID patent clusters, white spaces, key players, FTO risks. Produces strategic assessment → R&D direction, licensing, IP filing strategy.
Use When
- Before R&D in new tech area (what's claimed?)
- Market entry vs incumbents w/ strong portfolios
- Investment due diligence (IP asset assessment)
- Inform patent filing strategy (where, what to claim)
- FTO risk for new product/feature
- Monitor competitor IP for strategic positioning
In
- Required: Tech domain/product area
- Required: Geographic scope (US, EU, global)
- Optional: Specific competitors
- Optional: Own patent portfolio (gap analysis + FTO)
- Optional: Time horizon (5 years, 10 years, all)
- Optional: Classification codes (IPC, CPC)
Do
Step 1: Define Search Scope
Establish analysis boundaries.
- Tech domain precisely:
- Core (e.g., "transformer-based language models" not "AI")
- Adjacent (e.g., "attention mechanisms, tokenization, inference optimization")
- Exclude (e.g., "computer vision transformers" if focusing on NLP)
- Relevant classification codes:
- IPC (International Patent Classification) — broad, worldwide
- CPC (Cooperative Patent Classification) — more specific, US/EU std
- Search WIPO's IPC publication or USPTO's CPC browser
- Geographic scope:
- US (USPTO), EU (EPO), WIPO (PCT), specific national offices
- Most analyses start US + EU + PCT → broad coverage
- Time window:
- Recent: 3-5 years (current competitive)
- Full history: 10-20 years (mature areas)
- Watch expired patents opening design space
- Doc scope as Landscape Charter
→ Clear bounded scope specific enough actionable + broad enough captures competitive. Classification codes ID'd for systematic search.
If err: Too broad (thousands of results) → narrow via technical specificity or focus application area. Too narrow (few) → broaden adjacent tech. Right scope typically 100-1000 families.
Step 2: Harvest Patent Data
Collect data within scope.
- Query databases via Charter:
- Free: Google Patents, USPTO PatFT/AppFT, Espacenet, WIPO Patentscope
- Commercial: Orbit, PatSnap, Derwent, Lens.org (freemium)
- Combine keyword + classification codes → best coverage
- Build queries systematically:
Query Construction:
+-------------------+------------------------------------------+
| Component | Example |
+-------------------+------------------------------------------+
| Core keywords | "language model" OR "LLM" OR "GPT" |
| Technical terms | "attention mechanism" OR "transformer" |
| Classification | CPC: G06F40/*, G06N3/08 |
| Date range | filed:2019-2024 |
| Assignee filter | (optional) specific companies |
+-------------------+------------------------------------------+
- Download structured (CSV, JSON):
- Patent/app num, title, abstract, filing date
- Assignee/applicant, inventor(s)
- Classification codes, citation data
- Legal status (granted, pending, expired, abandoned)
- Dedup by family (group national filings of same invention)
- Record total family count + src databases
→ Structured dataset of families within scope, deduped + timestamped. Foundation for all subsequent.
If err: DB access limited → Google Patents + Lens.org (free) good coverage. Query returns too many (>5000) → add technical specificity. Too few (<50) → broaden keywords or add classification.
Step 3: Analyze Landscape
Map clusters, key players, trends.
- Cluster analysis: Group by sub-tech:
- Classification codes or keyword clustering → 5-10 sub-areas
- Count families per cluster
- ID growing (recent surges) vs mature (flat/declining)
- Key player analysis: Top 10 assignees by:
- Total family count (portfolio breadth)
- Recent filing rate (last 3 years — current activity)
- Avg citation count (quality proxy)
- Geographic breadth (US-only vs global)
- Trend analysis: Filing trends over window:
- Overall volume by year
- Volume by cluster by year
- New entrants (assignees filing 1st time in domain)
- Citation network: Most-cited patents (foundational):
- High forward citations = heavily relied upon
- Likely blocking patents or essential prior art
- Produce Landscape Map: clusters, players, trends, key patents
→ Clear picture: who owns what, where activity concentrated, how landscape evolving. Key blocking patents ID'd. White spaces visible.
If err: Dataset too small for meaningful clustering → combine into broader groups. 1 assignee dominates (>50%) → analyze portfolio as separate sub-landscape.
Step 4: ID White Spaces + Risks
Strategic insights from landscape.
- White space analysis (opportunities):
- Areas within scope w/ few/no filings
- Expired families → design space reopened
- Active areas w/ only 1 player (first-mover but no competition)
- White spaces adjacent to growing clusters (next frontier)
- FTO risk screening (threats) — adapted from
healtriage:- Critical: Granted directly covering planned product/feature
- High: Pending apps likely to grant relevant claims
- Medium: Granted in adjacent areas could be broadly interpreted
- Low: Expired, narrow claims, geographically irrelevant
- Competitive positioning:
- Where portfolio sits rel to competitors?
- Which competitors have blocking positions in target areas?
- Which interested in cross-licensing?
- Produce Strategic Assessment: white spaces, FTO risks, positioning, recs
→ Actionable strategic recs: where to file, what to avoid, who to watch, what risks need detailed FTO.
If err: FTO risks ID'd — preliminary ONLY — does NOT replace formal FTO opinion from patent attorney. Flag critical for legal review. White spaces seem too good → verify search scope not accidentally excluded.
Step 5: Doc + Recommend
Package assessment for decision-makers.
- Landscape Report sections:
- Exec summary (1 page: key findings, top risks, main recs)
- Scope + methodology (search terms, DBs, date range)
- Landscape overview (clusters, trends, key players + viz)
- White space analysis (opportunities ranked by strategic value)
- Risk assessment (FTO concerns ranked by severity)
- Recs (filing strategy, licensing targets, monitoring alerts)
- Supporting data:
- Family list (structured, sortable)
- Cluster map (visual)
- Filing trend charts
- Key patent summaries (top 10-20 most relevant)
- Ongoing monitoring:
- Alert queries for new filings in critical areas
- Review cadence (quarterly active, annually stable)
→ Complete landscape report → strategic IP decisions. Evidence-based, clearly scoped, actionable.
If err: Report too large → exec summary first + offer detailed sections on request. Exec summary always stands alone as decision doc.
Check
- Landscape Charter defines scope, classification, geography, time window
- Patent dataset harvested from multi DBs + deduped
- Clusters ID'd w/ filing counts + trend direction
- Top 10 key players profiled w/ portfolio metrics
- White spaces ID'd + ranked by strategic value
- FTO risks screened + classified by severity
- Key blocking patents ID'd w/ citation analysis
- Recs specific + actionable
- Limitations acknowledged (screening vs formal FTO opinion)
- Monitoring alerts defined for ongoing tracking
Traps
- Too broad scope: "AI patents" not landscape — ocean. Be specific about tech + app
- Single-DB reliance: No single DB complete coverage. ≥2 srcs
- Ignore families: Count individual filings vs families inflates. 1 invention in 10 countries = 1 family not 10
- Confuse apps w/ grants: Pending app not enforceable. Distinguish granted vs published apps
- White space misinterp: Empty area = "nobody tried" or "everybody tried + failed." Investigate before assuming opportunity
- Landscape as legal opinion: Strategic intelligence, not legal advice. FTO risks flagged need formal analysis by patent counsel
→
search-prior-art— detailed prior art search for specific inventions or patent validityscreen-trademark— trademark conflict screening + distinctiveness analysis for trademark sidefile-trademark— trademark filing procedures EUIPO, USPTO, Madrid Protocolsecurity-audit-codebase— risk assessment methodology parallels IP risk screeningreview-research— literature review skills apply to prior art analysis
Dépôt GitHub
Compétences associées
executing-plans
DesignUtilisez la compétence executing-plans lorsque vous disposez d'un plan de mise en œuvre complet à exécuter par lots contrôlés avec des points de contrôle de revue. Elle charge et examine le plan de manière critique, puis exécute les tâches par petits lots (3 tâches par défaut) tout en rapportant la progression entre chaque lot pour une revue par l'architecte. Cela garantit une mise en œuvre systématique avec des points de contrôle de qualité intégrés.
requesting-code-review
DesignCette compétence délègue un sous-agent réviseur de code pour analyser les modifications apportées au code par rapport aux exigences avant de poursuivre. Elle doit être utilisée après avoir terminé des tâches, implémenté des fonctionnalités majeures, ou avant une fusion vers la branche principale. La revue aide à détecter précocement les problèmes en comparant l'implémentation actuelle avec le plan initial.
connect-mcp-server
DesignCette compétence fournit un guide complet permettant aux développeurs de connecter des serveurs MCP à Claude Code via les transports HTTP, stdio ou SSE. Elle couvre l'installation, la configuration, l'authentification et la sécurité pour intégrer des services externes tels que GitHub, Notion et des API personnalisées. Utilisez-la lors de la configuration d'intégrations MCP, de la configuration d'outils externes ou du travail avec le Protocole de Contexte de Modèle de Claude.
web-cli-teleport
DesignCette compétence aide les développeurs à choisir entre les interfaces Web et CLI de Claude Code en fonction de l'analyse des tâches, puis permet une téléportation transparente des sessions entre ces environnements. Elle optimise le flux de travail en gérant l'état et le contexte de la session lors du passage entre le web, la CLI ou le mobile. Utilisez-la pour des projets complexes nécessitant différents outils à diverses étapes.
