Complete List of Allowed Tools for Patent-Disclosure-Skill Version 4.2.0

The patent-disclosure-skill version 4.2.0 provides 25+ officially supported command-line tools located in skills/patent-disclosure/tools/ that handle document conversion, CAD processing, mathematical formula rendering, and patent classification workflows.

The handsomestWei/patent-disclosure-skill repository (v4.2.0) bundles a comprehensive suite of standalone Python utilities designed to automate the patent disclosure workflow. These allowed tools function as the official command-line interface, covering everything from raw CAD data acquisition and structure validation to final document generation in Microsoft Word format.

Document Generation and Conversion Tools

The skill includes a complete pipeline for converting between Markdown, Word, and PowerPoint formats. These tools ensure that invention sketches and technical disclosures can flow seamlessly between editing environments.

md_to_docx.py renders Markdown-based disclosure drafts into Microsoft Word (.docx) files while preserving headings, tables, and image embeds. This is the primary output generator for final patent submissions.

python -m skills.patent-disclosure.tools.md_to_docx input.md output.docx

docx_to_md.py performs the reverse operation, extracting Word disclosure drafts back into Markdown for version-control editing. [pptx_to_md.py](https://github.com/handsomestWei/patent-disclosure-skill/blob/main/pptx_to_md.py) converts PowerPoint presentation slides (commonly used for invention sketches) into Markdown sections ready for the disclosure note.

For mathematical content, math_to_omml.py translates LaTeX math expressions into Office Math Markup Language (OMML) so they render correctly inside generated Word documents. When OMML is not desired, math_render.py renders LaTeX blocks as SVG images. The latex_delimiters.py utility normalizes LaTeX delimiters ($…$, \(...\), \[...\]) before feeding them to other math-related tools.

CAD Processing and Structure Visualization Tools

Version 4.2.0 includes specialized utilities for handling patent drawings extracted from CAD files. These tools ensure that line-art meets formal patent office requirements.

structure_lineart_gate.py filters and validates line-art drawings extracted from CAD files, ensuring they meet the formal requirements for "structure" sections. structure_lineart_compose.py merges multiple line-art assets into a single composite figure, handling layout, scaling, and labeling. For annotations, structure_callout_overlay.py generates call-out overlays (numbered boxes and arrows) on structure drawings, easing the reference between description and illustration.

design_lineart_gate.py performs similar validation specifically for design patents, checking ornamental drawings against patent-office format rules.

The CAD infrastructure tools include cad_scan.py, which scans directories of CAD files building a manifest of available components, and cad_formats.py, which maps supported file extensions (STEP, IGES, STL) to their respective import pipelines. To manage dependencies, cad_venv.py and bootstrap_cad_venv.py create isolated Python virtual environments containing heavy CAD-processing dependencies like OpenCascade.

Patent Classification and Data Acquisition Tools

patent_type.py determines the patent type (invention, utility model, or design) from a publication number or metadata—a prerequisite for downstream templating. This logic is implemented in skills/patent-disclosure/tools/patent_type.py and serves as the entry point for automated patent categorization.

browser.py provides a minimal headless-browser wrapper (based on Selenium) used to fetch remote patent PDFs or supplementary material. For diagram capture, svg_screenshot.py captures SVG diagrams as PNG/JPEG screenshots, useful for embedding visualized claim trees or flowcharts in the final disclosure.

The workflow orchestration tools include step_to_views.py, which orchestrates the step-wise generation of "views" (e.g., CAD extracts, annotated figures) for a given patent application, and run_step_to_views.py, a wrapper script that runs the entire step-to-views pipeline in one call while handling intermediate artifacts automatically.

Formula and Chemical Notation Processing

For patents containing chemical or mathematical formulas, version 4.2.0 provides a specialized processing chain.

formula_chem.py parses chemical notation (SMILES, InChI) feeding into the CAD-generation workflow. formula_units.py parses chemical/physical units within formulas, normalizing them for downstream rendering. The formula_paradigms.py tool provides a catalogue of common formula patterns (e.g., reaction equations) to aid automatic detection.

For safety and validation, formula_eval.py safely evaluates arithmetic parts of formulas (such as calculating stoichiometric coefficients) without executing arbitrary code. check_formula_plan.py verifies that planned formula extraction steps are feasible given the current input assets.

Workflow and Infrastructure Utilities

Supporting the main pipeline are several critical infrastructure tools. stdio_utf8.py ensures that all standard I/O streams use UTF-8 encoding—required for correct handling of Chinese characters and special symbols—and is imported automatically by most tools.

iteration_dialog_log.py generates a log of the interactive dialogue between the user and the skill, useful for auditing and reproducibility. image_gen.py serves as a general-purpose image generation helper (e.g., QR codes, placeholders) used by several other tools. For documentation purposes, gen_demo_snap_step.py produces a demonstrative "snapshot" of a single pipeline step.

Dependencies are managed through requirements-step.txt, which lists the Python packages that must be installed before any step-wise tool can run:

pip install -r skills/patent-disclosure/tools/requirements-step.txt

Summary

• The patent-disclosure-skill v4.2.0 officially supports 25+ command-line tools located in skills/patent-disclosure/tools/.
• Document conversion is handled by md_to_docx.py, docx_to_md.py, and pptx_to_md.py, with mathematical support via math_to_omml.py and math_render.py.
• CAD and structure visualization tools include structure_lineart_gate.py, structure_lineart_compose.py, and cad_scan.py for processing patent drawings.
• Patent classification relies on patent_type.py, while browser.py enables remote PDF acquisition.
• Chemical and mathematical formulas are processed by formula_chem.py, formula_eval.py, and related utilities.
• All tools are documented in the repository's SKILL.md and individual README.md files, with dependencies specified in requirements-step.txt.

Frequently Asked Questions

How do I invoke the allowed tools in patent-disclosure-skill v4.2.0?

Each tool is a standalone Python script invoked via the module execution syntax. For example, use python -m skills.patent-disclosure.tools.patent_type to determine patent classification, or python -m skills.patent-disclosure.tools.md_to_docx to convert Markdown to Word format. The full command structure for each tool is documented in skills/patent-disclosure/tools/README.md.

Which tool converts Markdown patent disclosures to Microsoft Word format?

md_to_docx.py performs this conversion while preserving document structure. Located in skills/patent-disclosure/tools/md_to_docx.py, it renders the Markdown-based disclosure draft into a .docx file, maintaining headings, tables, and embedded images. The reverse operation uses docx_to_md.py to extract Word documents back into Markdown for version control.

Are there tools for handling CAD files in patent disclosures?

Yes. cad_scan.py builds manifests of CAD components, while structure_lineart_gate.py validates line-art drawings against formal requirements. For design patents specifically, design_lineart_gate.py checks ornamental drawings against patent-office rules. These tools support STEP, IGES, and STL formats as mapped by cad_formats.py.

How does the skill handle mathematical formulas in generated documents?

The skill provides dual pathways: math_to_omml.py converts LaTeX expressions to Office Math Markup Language for native Word rendering, while math_render.py generates SVG image fallbacks. The latex_delimiters.py utility standardizes LaTeX syntax before processing, ensuring consistent handling of mathematical content across different document stages.

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