patent-analysis-guide
安装方式
先用 Prompt 核对来源,也可切换到安装命令或下载可用的安装包。
用 AI 助手安装
复制提示词,让助手先核对固定版本和依赖。
查看 Prompt收起 Prompt
先检查当前 Agent 是否已安装 SkillPI Skill Finder。如果尚未安装,请按官方仓库的说明安装并启用:https://github.com/Findskills/skillpi。然后使用 Skill Finder 打开此 SkillPI 详情页并定位 "patent-analysis-guide":https://skillpi.com/zh/skills/deepskill-92a50ec4aa58ee2a73f28ababdb38ec254f1dabc57b1adf9f9836db4cd0ade79。安装前核对上游来源和版本:https://github.com/brycewang-stanford/Auto-Empirical-Research-Skills/blob/bfa0a6e5a843babbf45a1ef1663fd5dc4f07c0ba/skills/43-wentorai-research-plugins/skills/domains/law/patent-analysis-guide/SKILL.md。来源不一致时停止并说明差异。核对通过后,用 Skill Finder 安装目标 Skill,并报告安装路径和来源版本。
此命令从 GitHub 来源仓库获取指定技能。运行前请先查看源文件。
npx skills add https://github.com/brycewang-stanford/Auto-Empirical-Research-Skills/tree/bfa0a6e5a843babbf45a1ef1663fd5dc4f07c0ba/skills/43-wentorai-research-plugins/skills/domains/law/patent-analysis-guide --skill patent-analysis-guide来源与配置
查看 GitHub 源码brycewang-stanford/Auto-Empirical-Research-Skills固定提交 bfa0a6e5a843babbf45a1ef1663fd5dc4f07c0ba
配置方式待核实
请查看原始来源中的适用工具、依赖和权限要求。
安装实测待补充
本条目面向法律事务场景,提供“patent-analysis-guide”技能的目录入口。公开标签:learning、专利检索、专利分类、专利分析。请通过固定上游版本核对实际功能、依赖与权限;此页为 Skillπ 独立编写的用途介绍和原站链接,不代表运行安全保证。原文与附件依各自许可开放。
文件浏览
---
name: patent-analysis-guide
description: "Patent search, classification, landscape analysis, and prior art mining"
metadata:
openclaw:
emoji: "📃"
category: "domains"
subcategory: "law"
keywords: ["patent", "intellectual-property", "prior-art", "cpc", "patent-landscape", "citation-analysis"]
source: "wentor"
---
# Patent Analysis Guide
A skill for conducting patent research, landscape analysis, and prior art searches. Covers patent database APIs, classification systems, citation network analysis, claim parsing, and technology trend mapping for intellectual property research.
## Patent Data Sources
### Major Patent Databases
| Database | Coverage | API | Cost |
|----------|----------|-----|------|
| USPTO PatentsView | US patents and applications | REST API, bulk download | Free |
| EPO Open Patent Services | EP, WO, and 100+ offices | REST API (OPS) | Free (throttled) |
| Google Patents | 120M+ documents worldwide | BigQuery (Google Patents Public) | Free (BigQuery costs) |
| Lens.org | 130M+ patent records | REST API | Free for researchers |
| WIPO PATENTSCOPE | PCT applications + national | REST API | Free |
### Programmatic Patent Search
```python
import requests
import xml.etree.ElementTree as ET
class EPOClient:
"""Client for the EPO Open Patent Services (OPS) API."""
BASE_URL = "https://ops.epo.org/3.2/rest-services"
def __init__(self, consumer_key: str, consumer_secret: str):
self.token = self._authenticate(consumer_key, consumer_secret)
def _authenticate(self, key: str, secret: str) -> str:
import base64
credentials = base64.b64encode(f"{key}:{secret}".encode()).decode()
resp = requests.post(
"https://ops.epo.org/3.2/auth/accesstoken",
headers={"Authorization": f"Basic {credentials}"},
data={"grant_type": "client_credentials"},
)
return resp.json()["access_token"]
def search(self, cql_query: str, max_results: int = 25) -> list[dict]:
"""
Search patents using CQL (Common Query Language).
Example queries:
ta="machine learning" AND cl="neural network"
pa="university" AND pd>=2020
"""
resp = requests.get(
f"{self.BASE_URL}/published-data/search",
headers={"Authorization": f"Bearer {self.token}",
"Accept": "application/json"},
params={"q": cql_query, "Range": f"1-{max_results}"},
)
return resp.json()
```
## Patent Classification Systems
### Cooperative Patent Classification (CPC)
The CPC hierarchy has five levels: Section > Class > Subclass > Group > Subgroup.
```
Example: H04L 9/3247
H = Electricity (Section)
H04 = Electric communication technique (Class)
H04L = Transmission of digital information (Subclass)
H04L 9/ = Cryptographic mechanisms (Group)
H04L 9/3247 = Digital signatures (Subgroup)
```
### IPC to CPC Mapping
```python
def parse_cpc_code(code: str) -> dict:
"""Parse a CPC classification code into its hierarchical components."""
code = code.strip().replace(" ", "")
return {
"section": code[0],
"class": code[:3],
"subclass": code[:4],
"group": code.split("/")[0] if "/" in code else code[:4],
"subgroup": code if "/" in code else None,
"full": code,
}
# Technology domain mapping (top-level CPC sections)
CPC_SECTIONS = {
"A": "Human Necessities",
"B": "Performing Operations; Transporting",
"C": "Chemistry; Metallurgy",
"D": "Textiles; Paper",
"E": "Fixed Constructions",
"F": "Mechanical Engineering; Lighting; Heating",
"G": "Physics",
"H": "Electricity",
"Y": "General Tagging of New Technological Developments",
}
```
## Patent Landscape Analysis
### Building a Patent Landscape
A patent landscape maps the technology and competitive environment in a domain:
```python
import pandas as pd
import numpy as np
from collections import Counter
def patent_landscape_metrics(patents: pd.DataFrame) -> dict:
"""
Compute patent landscape metrics from a patent dataset.
Expected columns: patent_id, filing_date, grant_date,
assignee, cpc_codes (list), claims_count, citations_received
"""
# Filing trend (annual)
patents["filing_year"] = pd.to_datetime(patents.filing_date).dt.year
annual_filings = patents.groupby("filing_year").size()
# Top assignees
top_assignees = patents.assignee.value_counts().head(20)
# Technology distribution (CPC subclass level)
all_cpc = []
for codes in patents.cpc_codes:
all_cpc.extend([c[:4] for c in codes])
cpc_distribution = Counter(all_cpc).most_common(20)
# Citation impact
citation_stats = patents.citations_received.describe()
# Geographic distribution (from assignee country)
geo_dist = patents.assignee_country.value_counts()
return {
"total_patents": len(patents),
"annual_filings": annual_filings.to_dict(),
"top_assignees": top_assignees.to_dict(),
"technology_areas": cpc_distribution,
"citation_stats": citation_stats.to_dict(),
"geographic_distribution": geo_dist.head(10).to_dict(),
}
```
### Citation Network Analysis
```python
import networkx as nx
def build_citation_network(patents: pd.DataFrame,
citations: pd.DataFrame) -> nx.DiGraph:
"""
Build a patent citation network.
citations: DataFrame with columns [citing_patent, cited_patent]
"""
G = nx.DiGraph()
# Add patent nodes with attributes
for _, row in patents.iterrows():
G.add_node(row.patent_id, assignee=row.assignee,
year=row.filing_year, cpc=row.cpc_codes[0][:4])
# Add citation edges
for _, row in citations.iterrows():
if row.citing_patent in G and row.cited_patent in G:
G.add_edge(row.citing_patent, row.cited_patent)
return G
def identify_seminal_patents(G: nx.DiGraph, top_n: int = 20) -> list:
"""Find the most influential patents by various centrality measures."""
in_degree = dict(G.in_degree())
pagerank = nx.pagerank(G)
# Combine metrics
scores = {}
for node in G.nodes():
scores[node] = {
"citations_received": in_degree[node],
"pagerank": pagerank[node],
}
ranked = sorted(scores.items(), key=lambda x: x[1]["pagerank"], reverse=True)
return ranked[:top_n]
```
## Claim Analysis
### Parsing Patent Claims
Patent claims define the legal scope of protection. Independent claims are the broadest; dependent claims narrow them:
```python
def parse_claims(claims_text: str) -> list[dict]:
"""
Parse patent claims text into structured claim objects.
Identifies independent vs dependent claims and extracts dependencies.
"""
# Split on claim numbers
claim_pattern = re.compile(r"\n\s*(\d+)\.\s+", re.MULTILINE)
parts = claim_pattern.split(claims_text)
claims = []
for i in range(1, len(parts), 2):
claim_num = int(parts[i])
claim_text = parts[i + 1].strip()
# Detect dependency
dep_match = re.match(
r"(?:The|A)\s+\w+\s+(?:of|according to)\s+claim\s+(\d+)",
claim_text, re.IGNORECASE
)
is_independent = dep_match is None
depends_on = int(dep_match.group(1)) if dep_match else None
claims.append({
"number": claim_num,
"text": claim_text,
"independent": is_independent,
"depends_on": depends_on,
"word_count": len(claim_text.split()),
})
return claims
```
## Prior Art Search Strategy
Systematic prior art search methodology:
1. **Define the invention**: Break the invention into key technical features
2. **Keyword search**: Use synonyms, broader terms, and technical variants
3. **Classification search**: Identify relevant CPC/IPC codes and search within them
4. **Citation search**: Forward and backward citation tracking from known relevant patents
5. **Assignee search**: Search patents from known competitors and research groups
6. **Non-patent literature**: Check academic papers, standards, product documentation
## Tools and Resources
- **PatentsView API**: Free US patent data with assignee disambiguation
- **Google Patents**: Full-text search with CPC browsing and citation links
- **Lens.org**: Scholarly and patent search with linking between patents and papers
- **Derwent Innovation**: Commercial tool for comprehensive patent analytics
- **PatSnap**: AI-powered patent intelligence platform
- **WIPO Pearl**: Multilingual patent terminology database
版本历史
-
v2026.07
GitHub ↗
原版更新说明
## Highlights - **First tagged release.** Every gate is green: `make check` (validate + 183 unit tests + eval-harness + numeric benchmarks), validate-catalog, quality-evals, and OpenSSF Scorecard on `main`. - **Rigor double loop at 13 method families.** Each family closes taxonomy tag → eval scenario (does the agent avoid the method's known traps?) → deterministic numeric benchmark (does it recover the true effect?). New this cycle: heterogeneous treatment effects (CATE) and quantile treatment effects (QTE). - **Paper-WorkFlow flagship at 29/29 executable rigor gates**, now merged to its `main` with a drift-gated README badge — the credibility artifact is checked by CI, so it can never silently go stale ([RIGOR.md](https://github.com/brycewang-stanford/Paper-WorkFlow/blob/main/RIGOR.md)). - **1,150 vendored skills / 69 collections**, all license-audited, security-scanned, tracked in a machine-readable catalog with CI-enforced freshness, and importable as a single whole-repo router skill. - **Hardened weekly upstream syncs**: synced trees must pass the full gate before merge; this cycle both sync PRs were verified, patched (StatsPAI's SkillOpt execution-gate card restored), and merged. - **First-party plugin marketplace** (Python / R / Stata empirical-analysis plugins + guide) shipped in the previous cycle, now under a tagged version. ## Known Follow-Ups - Ecosystem benchmark comparison (AERS vs econometrics agents on the shared benchmark suite) — [PLAN-2026-07 week 3](https://github.com/brycewang-stanford/Auto-Empirical-Research-Skills/blob/main/docs/PLAN-2026-07.md). - Decomposition (Oaxaca–Blinder) and shift-share (Bartik) method families (weeks 3–4). - GitHub Pages catalog site, README demo screenshots, five-language README consistency script (week 4). ## Catalog - Top-level collections: **69** - `SKILL.md` files: **1150** - Cataloged agent tools: **335** ([`tools/CATALOG.md`](../tools/CATALOG.md)) ## Methodological rigor - Method families in the coverage map: **13** ([`RIGOR_COVERAGE.md`](RIGOR_COVERAGE.md)) - Eval scenarios: **26** (138 rubric items, 122 auto-checkable) ([`../eval-harness/`](../eval-harness/README.md)) - Benchmark tasks: **13** (70 gold checks, 58 required) ([`../benchmark/`](../benchmark/README.md)) ### Method families (skills tagged) - `iv`: 27 - `rdd`: 25 - `panel-fe`: 19 - `staggered-did`: 17 - `bayesian`: 13 - `did`: 10 - `event-study`: 10 - `matching`: 10 - `synthetic-control`: 10 - `cate`: 8 - `dml`: 8 - `survival`: 3 - `quantile`: 2 ### Eval scenarios by category - `causal-identification`: 14 - `citation-hygiene`: 1 - `reproducibility`: 2 - `research-integrity`: 1 - `runtime-safety`: 1 - `statistical-validity`: 2 - `writing-compliance`: 3 - `writing-style`: 2 ## License buckets - AGPL-3.0: 1 - Apache-2.0: 1 - CC-BY-NC-4.0 (non-commercial): 1 - CC-BY-SA-4.0: 1 - CC-BY-SA-4.0 (repository default): 4 - GPL-3.0: 1 - MIT: 35 - MIT Non-Commercial: 1 - Mixed (per-skill; see folder): 2 - UNKNOWN - check upstream: 22 ## Quality gates - `make check` — catalog/provenance/audit/eval freshness, repo hygiene, Paper-WorkFlow gates, stdlib unit tests, eval-harness lint + graded smoke, and the strict numeric benchmark (reference-gold drift fails the build). - External links and OpenSSF Scorecard run on their scheduled workflows.
当前收录版本
0.0.1
质量评测
满分 100。前四项来自 DeepSkill 收录的质量报告。
评分依据
覆盖专利检索、分类、态势分析和现有技术挖掘全流程
评分依据
触发词覆盖全面:patent, intellectual-property, prior-art, cpc, patent-landscape, citation-analysis
评分依据
Phase划分清晰:数据源→分类体系→景观分析→引用网络→权利要求→现有技术策略
评分依据
工具来源明确:引用USPTO、EPO、Google Patents、Lens.org等权威数据源
计算方式
取前四项原始分值的算术平均,四舍五入取整。可通过收录报告核对各项评分与说明,运行记录需单独查证。
权限与使用前检查
根据当前主文档自动整理,命中内容可能是示例或防护说明;不是实际权限声明,也未覆盖附件、外部依赖或运行行为。
-
账号、密钥或授权
使用专用、最小权限账号,不把密钥粘贴到公开内容或命令日志;核对接收数据的服务与租户。
主文档提及 · 第 45, 47, 48, 51 行
文档引用的域名或地址
这些地址可能用于文档链接或示例,不是已批准的联网清单。使用前核对实际目的地及发送的数据。
- ops.epo.org
安全评级与访问范围
risk · 静态证据核验,不代表运行安全保证。
使用前请由宿主核对并授权实际文件读写、外部服务和联网目的地;本站的核验不会授予任何运行权限。
用户讨论
还没有已核实的点评。