What problem does it solve? Turning a research paper, thesis, or codebase into a Chinese invention patent application is slow and error-prone: claim features lose traceability to source evidence, formulas get mangled in Word, and unsupported features slip into formal claims. This Skill enforces an evidence-ledger workflow so every claim feature maps back to a source ID before drafting. ## Core Features & Use Cases - Evidence-constrained drafting: Builds stable source IDs (paper text, equations, figures, code) and an evidence ledger, excluding unsupported features from formal claims. - Full patent package generation: Produces claims, specification, abstract, abstract figure, and a review draft as separate DOCX files with native editable Office Math equations and black-and-white flowcharts (SVG/PNG). - Technical disclosure mode: Mines patent points from project documents and code, runs CNIPA-oriented prior-art search paths, and generates timestamped Chinese technical disclosure (技术交底书) in Markdown and Word with Mermaid diagrams. - Use Case: Given a PDF paper on a defect-detection model, generate a complete Chinese patent draft where every claim feature traces to a page or equation, formulas render as editable Word equations, and validation scripts flag structural claim errors before delivery. ## Quick Start Convert the attached paper PDF into a Chinese invention patent disclosure and claim draft with evidence mapping for each claim feature.
