Patent Disclosure Skill Directory Structure: Complete Guide to the handsomestWei Repository

The handsomestWei/patent-disclosure-skill repository organizes patent automation capabilities into modular Skill packages under the skills/ directory, with separate subdirectories for open-access data handling, disclosure generation, exam policies, and search functionality, alongside documentation, scripts, and configuration files at the root level.

The handsomestWei/patent-disclosure-skill repository provides a modular framework for patent-related automation tasks, structured to support the Instag​it platform's skill-based architecture. Understanding its directory structure is essential for contributors and developers who want to extend its capabilities or integrate specific patent processing workflows. The codebase follows a clear separation between core skill implementations, utility tools, and project documentation.

Top-Level Directory Layout

The repository root contains five primary directories and several configuration files that define the project’s scope and entry points:

├─ docs/                    # Documentation assets and screenshots

├─ skills/                  # Core Skill implementations (patent workflows)

├─ scripts/                 # Development and maintenance utilities

├─ .github/                 # CI/CD workflow definitions

├─ .gitignore               # Git ignore patterns

├─ LICENSE                  # License file

├─ README.md                # Project overview and setup guide

├─ INSTALL.md               # Detailed installation instructions

├─ requirements.txt         # Python dependencies

└─ SKILL.md                 # Top-level skill descriptor for Instag​it

The docs/ directory stores static assets such as thanks.jpg and illustrative PNG files used by the README, while .github/workflows/ houses automation like update-star-history.yml for repository analytics.

The skills/ Directory: Core Package Architecture

The skills/ directory is the heart of the repository, containing five distinct patent-processing capabilities. Each sub-directory follows a consistent layout: a tools/ package containing reusable Python modules, plus a SKILL.md descriptor file that declares the skill’s interface to the Instag​it platform.

patent-oa: Open-Access Patent Data Handling

Located at skills/patent-oa/, this skill manages open-access patent data ingestion and vector search preparation. Its tools/ sub-directory includes:

• config.py – Loads YAML configuration settings including API keys and model parameters
• embed.py – Implements text embedding logic for vector search using OpenAI embeddings
• ingest_case.py – Handles the case-ingestion pipeline for patent documents

The skill is described in skills/patent-oa/SKILL.md, which defines its capabilities for the platform.

patent-disclosure: Disclosure Generation Workflow

The skills/patent-disclosure/ directory contains the rendering pipeline that transforms structured patent data into visual disclosures. Key utilities in its tools/ folder include:

• md_to_docx.py – Converts Markdown disclosure documents into Microsoft Word format
• pptx_to_md.py – Transforms PowerPoint presentations into Markdown for processing
• structure_lineart_gate.py – Implements gating logic for line-art processing in design patents

This skill also maintains its own SKILL.md descriptor at skills/patent-disclosure/SKILL.md.

patent-exam-policy: Policy References and Prompts

The skills/patent-exam-policy/ skill organizes regulatory and examination guidelines with:

• prompts/ – Template files for intake forms and guardrail constraints
• references/ – Topic-to-prompt mapping files and source lists

patent-reader and patent-search: UI and Retrieval

• patent-reader/ – Contains a simple patent-reading UI implementation with its own SKILL.md
• patent-search/ – Implements search-oriented functionality with prompts/patent_search.md and config.yaml for query configuration

Supporting Infrastructure

Beyond the core skills, the repository includes several support directories:

docs/ – Stores image assets including 效果例‑外观专利线稿.png (design patent line-art examples) and other illustrative screenshots referenced by documentation.

scripts/ – Contains maintenance utilities such as generate-star-history.py, which builds star-history charts for repository analytics.

.github/ – Hosts GitHub Actions workflows, specifically workflows/update-star-history.yml for automated repository metric updates.

Practical Code Examples

The following snippets demonstrate how to import and use utilities from the repository’s skill packages. Note that while directory names use hyphens, Python imports replace these with underscores.

Loading Configuration from patent-oa

from skills.patent_oa.tools.config import load_config

cfg = load_config()  # Reads skills/patent-oa/tools/config.yaml

print(cfg["openai_api_key"])  # Access specific configuration values

Source: skills/patent-oa/tools/config.py

Generating Text Embeddings for Vector Search

from skills.patent_oa.tools.embed import embed_text

text = "A method for producing a renewable polymer..."
vector = embed_text(text)  # Returns NumPy array for ANN search

print(vector.shape)  # → (768,)

Source: skills/patent-oa/tools/embed.py

Converting Markdown Disclosures to DOCX

from skills.patent_disclosure.tools.md_to_docx import md_to_docx

md_path = "output/disclosure.md"
docx_path = "output/disclosure.docx"

md_to_docx(md_path, docx_path)  # Generates styled Word document

Source: skills/patent-disclosure/tools/md_to_docx.py

Summary

• The skills/ directory contains five specialized packages (patent-oa, patent-disclosure, patent-exam-policy, patent-reader, patent-search) that implement distinct patent-related capabilities.
• Each skill follows a standardized structure with a tools/ sub-directory for Python utilities and a SKILL.md descriptor file for platform integration.
• The patent-oa skill handles data ingestion and vector embedding for search, while patent-disclosure manages document conversion and rendering pipelines.
• Supporting directories include docs/ for static assets, scripts/ for maintenance utilities, and .github/ for CI/CD workflows.
• Root-level files such as requirements.txt, INSTALL.md, and the top-level SKILL.md provide dependency management and high-level project metadata.

Frequently Asked Questions

What is the purpose of the SKILL.md files throughout the repository?

Each SKILL.md file serves as a metadata descriptor that declares the skill’s capabilities, inputs, and outputs to the Instag​it platform. According to the repository structure, these files exist at both the root level (defining the overall package) and within each skill sub-directory (defining specific patent-processing modules), enabling the platform to discover and invoke the correct tools automatically.

How does the patent-oa skill handle vector search embeddings?

The patent-oa skill implements embedding functionality in skills/patent-oa/tools/embed.py, which provides the embed_text() function. This utility converts patent text into high-dimensional vectors (typically 768 dimensions) suitable for approximate nearest neighbor (ANN) search, enabling semantic retrieval of similar patent documents from the vector store.

Can I use the markdown-to-DOCX converter independently of the full pipeline?

Yes. The md_to_docx.py module in skills/patent-disclosure/tools/ is designed as a standalone utility. You can import the md_to_docx() function directly, pass it a Markdown file path and an output DOCX path, and generate formatted Word documents without running the complete disclosure generation workflow, making it reusable for other documentation tasks.

Where are the CI/CD and automation configurations stored?

Continuous integration configurations are located in .github/workflows/, specifically the update-star-history.yml file which automates repository analytics. Additionally, the scripts/ directory at the repository root contains development utilities like generate-star-history.py, which supports maintenance tasks but is not part of the core patent-processing skills.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}