general-patent-search

SkillSearch

Search for patents by keyword, material name, or assignee using free data sources (Google Patents).

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the general-patent-search skill

What this skill tells your AI

The instructions your AI receives, as published by learningmatter-mit/atomisticskills in .agents/skills/general-patent-search/SKILL.md and read by ahel’s review.

Goal

To retrieve patent information (titles, assignees, dates, and snippets) for specific materials, chemicals, or technologies to assess novelty, freedom-to-operate, and existing intellectual property landscapes. This skill uses Google Patents for free, global, no-key-required patent querying.

Instructions

1. Identify Target and Scope

Determine exactly what you are searching for.

  • Are you looking for a specific chemical? (e.g., "tetrafluoropropene")
  • Are you looking for a specific company's IP? (e.g., assignee "SK Innovation")

2. Search Strategy (Google Patents)

Use the python script to query Google Patents. It will return global patents matching your query string.

# Env: base-agent
python .agents/skills/general-patent-search/scripts/query_google_patents.py "query string" --limit 10

Examples

Searching for recent global patents on a novel refrigerant by chemical name:

# Env: base-agent
python .agents/skills/general-patent-search/scripts/query_google_patents.py "tetrafluoropropene OR HFO-1234yf" --limit 5

Searching for global patents assigned to SK Innovation regarding electrolytes:

# Env: base-agent
python .agents/skills/general-patent-search/scripts/query_google_patents.py "assignee:(SK Innovation) AND electrolyte" --limit 5

Constraints

  • Environments: The script requires the base-agent Conda environment.
  • Google Patents Limits: The scraping script is sensitive to IP rate-limiting. Do not set --limit excessively high (keep under 50).
  • Chemical Structures: This tool performs text-based searches. It does not perform exact Substructure or Tanimoto similarity searches natively. You must provide textual synonyms (e.g., IUPAC names, trade names) for accurate chemical retrieval.

References


Author: Sathya Edamadaka Contact: GitHub @snme

Signals

GitHub stars
164
Forks
24
Last commit
Sep 2026
Advanced
Catalog kind
skill
Gateway key
general-patent-search
Source
github.com/learningmatter-mit/atomisticskills