Patent Disclosure Skill Schema Files: Complete Reference for YAML Data Contracts

The Patent Disclosure Skill repository uses 16 YAML-based schema files to define strict data contracts for patent disclosure, OA prosecution, docketing, and application workflows, ensuring validated, machine-readable output for SVG generation and vector database indexing.

The handsomestWei/patent-disclosure-skill repository implements a contract-first architecture where patent disclosure skill schema files govern every data interaction across the skill ecosystem. These immutable YAML definitions act as blueprints that downstream code validates against, guaranteeing consistent structure for mechanical inventions, industrial designs, and prosecution materials before they enter automated pipelines.

Core Disclosure Schemas: Structure, Appearance, and Assets

The foundation of the disclosure skill rests on three primary schemas that capture distinct aspects of an invention.

StructureSchema for Mechanical Inventions

Located at skills/patent-disclosure/references/schemas/structure.schema.yaml, StructureSchema defines the mechanical-structure description of utility models and inventions. It captures parts, relations, spatial information, functional rationale, and uncertainty flags for disclosed structures. Downstream validators use this schema to ensure that every mechanical component documented has required fields like id, name, and shape before Markdown or SVG generation proceeds.

AppearanceSchema for Industrial Designs

The AppearanceSchema at skills/patent-disclosure/references/schemas/appearance.schema.yaml handles visual-design descriptions for industrial-design patents. It records product name, overall shape, form classification (solid or planar), claimed faces, view selections, ornaments, colour specifications, design points, and uncertainty annotations. This schema ensures that visual assets align with legal requirements for design patent applications.

FigurePlanSchema as the Asset Catalogue

FigurePlanSchema (skills/patent-disclosure/references/schemas/figure_plan.schema.yaml) serves as the linking hub between structure and appearance schemas to actual drawing assets. It provides a catalogue of source images, figure IDs, and relationship mapping via fields like relates_to that other schemas reference for seen_in or source_image attributes. Without this central registry, the system cannot resolve which SVG files correspond to which structural parts or design views.

Line-Art Composition and Visualization Schemas

Four specialized schemas control how vector graphics are assembled, referenced, and annotated for patent drawings.

StructureLineartComposeSchema

The StructureLineartComposeSchema at skills/patent-disclosure/references/schemas/structure_lineart_compose.schema.yaml controls how individual parts assemble into composite SVG line-art. It describes each part’s SVG file path, composition groups, and restrictions on further subdivision—specifically enforcing rules such as no splitting of bolts into sub-IDs once defined.

StructureLineartBriefSchema

For quick reference during rendering, StructureLineartBriefSchema (skills/patent-disclosure/references/schemas/structure_lineart_brief.schema.yaml) stores a lightweight mapping of part IDs to SVG snippets and anchors. The "call-out" view generation step consumes this schema to rapidly locate graphic elements without parsing full composition trees.

StructureCalloutAnchorsSchema

StructureCalloutAnchorsSchema (skills/patent-disclosure/references/schemas/structure_callout_anchors.schema.yaml) persists normalized coordinates and base SVG references needed to attach call-out annotations to assembled line-art. These anchor points ensure that leader lines and reference numerals maintain spatial relationships when the underlying graphics scale or transform.

DesignLineartBriefSchema

The visual counterpart for appearance data, DesignLineartBriefSchema (skills/patent-disclosure/references/schemas/design_lineart_brief.schema.yaml), links the appearance schema to generated line-art files. It ensures that claimed faces and design points specified in appearance.schema.yaml correctly map to their visual representations in the output documents.

FormulaPlanSchema for Technical Documentation

FormulaPlanSchema (skills/patent-disclosure/references/schemas/formula_plan.schema.yaml) outlines how formulae—whether chemical equations or circuit diagrams—are documented. It provides fields for formula source materials, rendering hints, and uncertainty flags, ensuring that technical expressions receive the same validation treatment as structural components.

Prosecution and Administrative Workflow Schemas

Beyond invention disclosure, the repository manages patent prosecution through dedicated workflow schemas.

OA Playbook and Case Schemas

The patent-oa skill relies on two complementary schemas: oa_playbook.schema.yaml and oa_case.schema.yaml, both located under skills/patent-oa/references/schemas/. The playbook schema structures metadata for OA "playbooks" (collections of prosecution tactics), while the case schema defines individual OA cases that embed into the knowledge vault. Together, they enable systematic response generation for office actions.

ExamPolicyBacklogSchema

ExamPolicyBacklogSchema (skills/patent-exam-policy/references/schemas/exam_policy_backlog.schema.yaml) models the backlog of exam-policy items awaiting review. It captures policy-related entries with fields for status, priority, and notes, allowing the system to track evolving examination guidelines that might affect pending applications.

Patent Docket Schema

Located at skills/patent-docket/references/docket.schema.yaml, the docket schema organizes phases, dispositions, and case-specific metadata for patent family tracking. It defines the workflow states that drive deadline management and status reporting across the patent lifecycle.

Application Metadata and Registration Schemas

The final category handles application-level identifiers and planning documents.

NumeralRegisterSchema

NumeralRegisterSchema (skills/patent-application/references/schemas/numeral_register.schema.yaml) registers numeric identifiers used across disclosures. It aligns part IDs with human-readable names, ensuring that reference numeral "1" consistently maps to "外壳" (housing) or equivalent terms across all downstream schemas and generated documents.

InventionFiguresSchema

InventionFiguresSchema (skills/patent-application/references/schemas/invention_figures.schema.yaml) enumerates figures belonging to an invention, listing figure IDs, captions, and source images. This feed supports both structure and appearance schemas by providing the authoritative list of visual assets associated with a given disclosure.

ApplicationPlanSchema

ApplicationPlanSchema (skills/patent-application/references/schemas/application_plan.schema.yaml) provides a high-level plan for filing an application, describing filing strategy, target jurisdictions, and timelines. It serves as the coordinating document that triggers the generation of supporting schemas and validates that all required disclosure components are complete before submission.

Validating and Implementing Schema Contracts

Each skill reads these YAML files, validates them against the matching $schema keyword, and produces appropriate artifacts. The following Python pattern demonstrates loading and validating against structure.schema.yaml:

import yaml
import jsonschema

def load_schema(path):
    with open(path, encoding="utf-8") as f:
        return yaml.safe_load(f)

structure_schema = load_schema(
    "skills/patent-disclosure/references/schemas/structure.schema.yaml"
)

def validate_structure(doc):
    """Validate a user-provided structure document."""
    jsonschema.validate(instance=doc, schema=structure_schema)

Generating a compliant structure document requires specifying the schema reference and required fields:

$schema: structure.schema
version: 1
mode: disclosure
parts:
  - id: "1"
    name: "外壳"
    shape: "壳体"
    material_hint: "铝合金"
relations: []
spatial: []
function_of_structure: ["保护内部电子元件"]

To ensure consistency between appearance declarations and available figures, cross-schema validation links views to the figure plan:

appearance = yaml.safe_load(open("appearance.yaml"))
figure_plan = yaml.safe_load(open("figure_plan.yaml"))

# Ensure each declared view has a corresponding figure entry

for view in appearance["views"]:
    assert any(fig["id"] == view["source_image"] for fig in figure_plan["figures"])

Summary

• Patent disclosure skill schema files reside in four primary domains: disclosure (patent-disclosure), prosecution (patent-oa), policy (patent-exam-policy), and docketing (patent-docket).
• Core disclosure schemas include structure.schema.yaml for mechanical parts, appearance.schema.yaml for industrial designs, and figure_plan.schema.yaml as the central asset registry.
• SVG generation schemas (structure_lineart_compose.schema.yaml, structure_lineart_brief.schema.yaml, structure_callout_anchors.schema.yaml, and design_lineart_brief.schema.yaml) control vector graphic assembly and annotation.
• Prosecution schemas (oa_playbook.schema.yaml, oa_case.schema.yaml, exam_policy_backlog.schema.yaml, and docket.schema.yaml) manage office action workflows and administrative tracking.
• Application schemas (application_plan.schema.yaml, invention_figures.schema.yaml, and numeral_register.schema.yaml) coordinate filing strategies and maintain identifier consistency across documents.

Frequently Asked Questions

What is the purpose of FigurePlanSchema in patent disclosures?

FigurePlanSchema acts as the central catalogue linking structural and appearance data to physical drawing assets. It stores figure IDs, source image paths, and relates_to mappings that allow the system to resolve which SVG or raster files correspond to specific invention parts or design views.

How do StructureSchema and AppearanceSchema differ?

StructureSchema (structure.schema.yaml) describes mechanical utility inventions through parts, spatial relations, and functional rationales, while AppearanceSchema (appearance.schema.yaml) documents industrial designs through visual attributes like shape, form, claimed faces, and ornaments. The former targets utility models; the latter targets design patents.

Which schema files control SVG call-out generation?

Three schemas collaborate on call-outs: structure_lineart_compose.schema.yaml defines how parts assemble into composite images; structure_lineart_brief.schema.yaml provides rapid part-to-SVG mapping; and structure_callout_anchors.schema.yaml stores the normalized coordinates where annotation leader lines attach to the graphics.

Where are OA prosecution schemas located in the repository?

OA-related schemas reside under skills/patent-oa/references/schemas/, specifically oa_playbook.schema.yaml for tactic collections and oa_case.schema.yaml for individual case metadata. The exam-policy schema lives in skills/patent-exam-policy/references/schemas/exam_policy_backlog.schema.yaml.

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