prompt
FreeEnd-to-end China patent mining and technical disclosure drafting.
FreeFree tier
Inputs: text, codeOutputs: text
About prompt
A specialized prompt that acts as a patent-engineering agent, guiding an AI to mine patentable inventions from technical projects and draft production-ready Chinese patent disclosure documents (技术交底书). It follows CNIPA norms, de-identification rules, and prior-art discipline. The prompt covers intake boundary clarification, project scanning for patentable material, patent-point mining fusion, de-identified disclosure drafting with mermaid diagrams, iterative revision loops, and self-check gates. It is part of the ai-boost/awesome-prompts collection on GitHub.
Key Features
Intake boundary clarification: confirms technical theme, claim type inclination, and contact.
Project scanning for patentable material: reads patent docs, design docs, code, and architecture docs.
Patent-point mining fusion: lists 3–5 candidate points with background, innovation, and differentiation.
De-identified disclosure drafting: produces agent-structured disclosure packages for CNIPA.
Iterative revision loops and self-check gates for quality assurance.
Generates mermaid diagrams for technical illustrations.
Pros & Cons
Pros
- Specialized for Chinese patent norms and CNIPA requirements.
- Provides a structured, multi-step approach to patent mining and drafting.
- Includes de-identification and iterative revision loops for high-quality outputs.
- Open source and freely available on GitHub.
- Covers both method and system claim classes.
Cons
- Requires an AI model to execute; effectiveness depends on the underlying model.
- Does not substitute for legal advice or a licensed patent attorney.
- The prompt is complex and may need customization for specific project contexts.
- No built-in integration with patent databases for automated prior-art search.
Best For
Mining patentable inventions from software or hardware projects.Drafting Chinese patent disclosure documents for attorneys.Conducting CNIPA prior-art searches.De-identifying proprietary information while preserving technical details.Structuring technical disclosures for method and system patent filings.