The problem

Vague technical notes, snippets of code, or rough product ideas often fail the "sufficient disclosure" requirements of patent law. Patent agents frequently waste billable hours chasing engineers for missing mathematical definitions, prior art defects, or logical steps needed for a valid filing.

What it does

• Converts minimal inputs—like a single code snippet or a one-sentence idea—into a structured technical disclosure document.
• Enforces legal compliance by ensuring every mathematical symbol is defined in LaTeX and all technical problems map one-to-one to background defects.
• Generates bidirectional formulas and qualitative effect mechanisms that avoid the legal risks of fabricated performance data.
• Standardizes technical terminology and formats parameter tables and pseudocode to meet Patent Law Article 26 standards.

Why this beats prompting it yourself

Most LLMs fabricate specific percentages or use promotional language like "revolutionary" which can invalidate a patent claim. This skill is hardcoded with legal constraints to prioritize qualitative mechanisms over fake data, ensure symbol consistency, and force the inclusion of verifiable prior art citations.

Use cases

• Converting undocumented Python or C++ logic into a formal invention disclosure.
• Formalizing "emotional" product feedback into a technical system design for IP protection.
• Drafting the initial technical foundation for a patent agent to review, reducing back-and-forth cycles.

Known limitations

Does not generate actual vector diagrams or CAD files. Use placeholders for experimental data where actual numbers are required for filing.