Back to Details

code2patent

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

代码仓库转专利交付

Code Repository to Patent Delivery

定位

Positioning

本技能用于把已经开发完成的代码项目整理成专利代理师可继续起草和判断的材料。核心目标不是把代码翻译成专利语言，而是把真实实现、技术问题、技术方案、技术效果和证据位置整理成可追溯的发明专利底稿。
默认法域为中国发明专利，主要适用于软件、算法、Agent 系统、调度优化、状态表示、鉴权、记忆、上下文编排和文件系统等以代码实现为主的方案。
推荐主链路：
text
用户材料 + 代码仓库
        ↓
输入成熟度判断 + archive 归档目录
        ↓
项目边界与依赖画像
        ↓
方案-代码证据映射
        ↓
算法/软件类说明书式技术交底书
        ↓
权利要求布局卡 + 权利要求-证据矩阵
        ↓
发明专利初稿 + 初稿自检表
This skill is used to organize completed code projects into materials that patent attorneys can continue to draft and judge. The core goal is not to translate code into patent language, but to organize real implementations, technical problems, technical solutions, technical effects, and evidence locations into traceable invention patent drafts.
The default jurisdiction is Chinese invention patents, which are mainly applicable to software, algorithms, Agent systems, scheduling optimization, state representation, authentication, memory, context orchestration, file systems and other implementations mainly based on code.
Recommended main workflow:
text
User Materials + Code Repository
        ↓
Input Maturity Judgment + archive Directory
        ↓
Project Boundary and Dependency Profile
        ↓
Solution-Code Evidence Mapping
        ↓
Algorithm/Software Specification-Style Technical Disclosure Document
        ↓
Claim Layout Card + Claim-Evidence Matrix
        ↓
Invention Patent Draft + Draft Self-Check List

启动闸门

Start Gate

开始读取代码前，先确认：
  1. 是否已有候选专利方案清单。
  2. 候选清单是仅有名称，还是已有技术问题、核心实现、技术效果和代码范围。
  3. 本次目标是技术交底书、权利要求布局卡、发明专利初稿，还是候选方案挖掘。
  4. 是否有 PRD、需求说明、会议纪要、客户交底书模板、既有申请样式或代理师格式要求。
  5. 本次归档主题是什么，用于创建 
    archive/YYYY-MM-DD-主题/
    。
专利名称不是技术方案。 如果用户只给标题或一句话方向，不要直接写交底书或初稿。
Before starting to read the code, confirm the following:
  1. Is there already a list of candidate patent solutions?
  2. Does the candidate list only have names, or does it already include technical problems, core implementations, technical effects, and code scope?
  3. Is the goal of this task a technical disclosure document, claim layout card, invention patent draft, or candidate solution mining?
  4. Are there PRDs, requirement descriptions, meeting minutes, client disclosure templates, existing application styles, or format requirements from patent attorneys?
  5. What is the archiving theme for this task, which will be used to create 
    archive/YYYY-MM-DD-Theme/
    .
Patent titles are not technical solutions. If the user only provides a title or one-sentence direction, do not directly write a disclosure document or draft.

输入成熟度分流

Input Maturity Diversion

等级输入状态默认动作
T0 名称清单只有专利名称、标题或一句话方向输出《专利名称反向澄清卡》，等待用户确认
T1 方向清单有名称和业务方向，但缺少技术问题、实现路径或效果先补齐候选解释、代码证据方向和待确认问题
T2 技术方案清单已有技术问题、核心实现、目标效果和初步模块范围进入定向检索，生成代码证据映射和交底书
T3 证据型方案清单已有技术方案、代码路径、关键流程和证据等级复核证据后继续生成交底书、布局卡或初稿
无清单用户希望从代码中挖掘专利点先输出候选可专利方案清单，人工筛选后再起草
未经人工确认的 T0/T1 或自动挖掘结果，只能作为候选方向，不得写成客户已确认的技术方案。
LevelInput StatusDefault Action
T0 Name ListOnly patent names, titles, or one-sentence directionsOutput "Patent Title Reverse Clarification Card" and wait for user confirmation
T1 Direction ListHas names and business directions, but lacks technical problems, implementation paths, or effectsFirst supplement candidate explanations, code evidence directions, and pending confirmation questions
T2 Technical Solution ListAlready has technical problems, core implementations, target effects, and preliminary module scopeEnter targeted retrieval, generate code evidence mapping and disclosure documents
T3 Evidence-Based Solution ListAlready has technical solutions, code paths, key processes, and evidence levelsReview evidence and then continue to generate disclosure documents, layout cards, or drafts
No ListUser wants to mine patent points from codeFirst output a list of candidate patentable solutions, then draft after manual screening
Unconfirmed T0/T1 or automatically mined results can only be used as candidate directions, and cannot be written as customer-confirmed technical solutions.

材料优先级

Material Priority

执行时先读用户提供材料，再读代码。建议顺序：
  1. 候选专利方案清单或标题清单
  2. PRD、产品方案、需求说明
  3. 沟通纪要、录音转写、客户补充说明
  4. 技术交底书模板、既有申请文件、代理师格式要求
  5. 代码仓库 README、架构文档、目录说明
  6. 源代码、配置、测试、部署和运行时文件
如果用户提供模板或样本，先抽取章节结构、步骤编号、附图习惯和技术效果写法，再生成交付物。若未提供模板，按 
references/algorithm-software-disclosure-format.md
和内置模板执行。
When executing, read user-provided materials first, then read the code. Recommended order:
  1. Candidate patent solution list or title list
  2. PRD, product plan, requirement description
  3. Communication minutes, audio transcriptions, customer supplementary instructions
  4. Technical disclosure templates, existing application documents, patent attorney format requirements
  5. Code repository README, architecture documents, directory descriptions
  6. Source code, configuration, testing, deployment, and runtime files
If the user provides a template or sample, first extract the chapter structure, step numbers, drawing habits, and technical effect writing methods, then generate deliverables. If no template is provided, follow 
references/algorithm-software-disclosure-format.md
and built-in templates.

必读规范

Mandatory Specifications

只在需要时读取对应文件，避免把所有 reference 一次性装入上下文。
需要处理的问题读取文件
算法/软件类交底书结构、S1...Sn、代码证据后置规则
references/algorithm-software-disclosure-format.md
项目技术方案画像、依赖边界、自研与第三方能力区分
references/project-analysis-spec.md
S/F/E 抽取、A/B/C 分级、从代码到专利表达的转译
references/code-extraction-spec.md
快速进入布局卡、初稿和自检写作
references/patent-drafting-quick-reference.md
需要完整起草依据、权利要求层级和摘要规则
references/patent-drafting-spec.md
本文件只保留入口、分流、路由和强规则；细节规则以后优先维护在 
references/
。
Only read corresponding files when necessary, avoid loading all references into the context at once.
Problem to SolveFile to Read
Algorithm/software disclosure structure, S1...Sn, code evidence post-positioning rules
references/algorithm-software-disclosure-format.md
Project technical solution profile, dependency boundaries, distinction between self-developed and third-party capabilities
references/project-analysis-spec.md
S/F/E extraction, A/B/C grading, translation from code to patent expression
references/code-extraction-spec.md
Quick access to layout card, draft, and self-check writing
references/patent-drafting-quick-reference.md
Complete drafting basis, claim hierarchy, and abstract rules required
references/patent-drafting-spec.md
This file only retains entry points, diversion, routing, and strong rules; detailed rules will be prioritized in 
references/
in the future.

模板路由

Template Routing

目标产物模板
专利名称反向澄清卡
templates/patent-title-clarification-card-template.md
算法/软件类发明专利技术交底书
templates/invention-patent-disclosure-template.md
权利要求布局卡
templates/invention-patent-claim-layout-template.md
权利要求-证据矩阵
templates/invention-patent-claim-evidence-matrix-template.md
发明专利初稿
templates/invention-patent-draft-template.md
发明专利初稿自检表
templates/invention-patent-draft-self-check-template.md
templates/
是可直接填充的交付骨架；
references/
是执行规则和判断标准。不要把二者合并。
Target DeliverableTemplate
Patent Title Reverse Clarification Card
templates/patent-title-clarification-card-template.md
Algorithm/Software Invention Patent Technical Disclosure Document
templates/invention-patent-disclosure-template.md
Claim Layout Card
templates/invention-patent-claim-layout-template.md
Claim-Evidence Matrix
templates/invention-patent-claim-evidence-matrix-template.md
Invention Patent Draft
templates/invention-patent-draft-template.md
Invention Patent Draft Self-Check List
templates/invention-patent-draft-self-check-template.md
templates/
is a directly fillable delivery skeleton; 
references/
is the execution rules and judgment criteria. Do not merge the two.

核心强规则

Core Strong Rules

  1. 先判断输入成熟度，再读代码细节。
  2. 用户提供的方案、模板和样本优先于 AI 自行推断。
  3. 没有人工确认的候选方向，不直接进入交底书或专利初稿。
  4. 先做项目边界与依赖画像，避免把第三方库、模型平台或框架默认能力写成申请人的创新。
  5. 代码证据映射和技术交底书正文必须分层。
  6. 交底书正文默认采用算法/软件类说明书式结构，不写成代码盘点报告。
  7. 文件路径、函数名、字段名、测试文件和内部事件名默认进入证据映射或附录；正文只保留必要证据摘要。
  8. 发明内容优先整理为技术问题、总体方案、步骤 S1...Sn、进一步限定和技术效果。
  9. 具体实施方式用示例场景展开输入、模型/状态、步骤、参数、输出和替代实现。
  10. 代码架构必须理解，但最终正文应转译为技术对象、关系、状态、动作、时序和输出，不展示技术选型或模块清单。
  11. 05-技术交底书
    、
    05A-权利要求布局卡
    、
    05B-权利要求-证据矩阵
    、
    06-发明专利初稿
    中的 S1...Sn 步骤编号应保持一致。
  12. 目标是初稿时，默认先生成布局卡和证据矩阵，再生成完整初稿。
  13. A 级证据可进入独立权利要求骨架；B 级证据优先进入从属权利要求或优选实施例；C 级证据只能进入待确认事项或替代实施例。
  14. 摘要默认控制在 300 字内，不使用商业性宣传语言。
  15. 正式申请文件仍需专利代理师审核。
  1. Judge input maturity first, then read code details.
  2. User-provided solutions, templates, and samples take precedence over AI's own inferences.
  3. Candidate directions without manual confirmation cannot directly enter the disclosure document or patent draft.
  4. First create a project boundary and dependency profile to avoid writing default capabilities of third-party libraries, model platforms, or frameworks as the applicant's innovation.
  5. Code evidence mapping and technical disclosure document body must be layered.
  6. The default structure of the disclosure document body is algorithm/software specification-style, not a code inventory report.
  7. File paths, function names, field names, test files, and internal event names are by default included in evidence mapping or appendices; only necessary evidence summaries are retained in the body.
  8. The invention content is prioritized to be organized into technical problems, overall solutions, steps S1...Sn, further limitations, and technical effects.
  9. Specific implementation methods use example scenarios to expand input, model/state, steps, parameters, output, and alternative implementations.
  10. The code architecture must be understood, but the final body should be translated into technical objects, relationships, states, actions, sequences, and outputs, without displaying technical selections or module lists.
  11. The step numbers S1...Sn in 
    05-Technical Disclosure Document
    , 
    05A-Claim Layout Card
    , 
    05B-Claim-Evidence Matrix
    , and 
    06-Invention Patent Draft
    must be consistent.
  12. When the target is a draft, first generate the layout card and evidence matrix by default, then generate the complete draft.
  13. Level A evidence can enter the independent claim framework; Level B evidence is prioritized for dependent claims or preferred embodiments; Level C evidence can only enter pending confirmation items or alternative embodiments.
  14. The abstract is controlled within 300 words by default, without using commercial promotional language.
  15. Formal application documents still require review by patent attorneys.

输出层级

Output Levels

层级适用场景主要产物
L0 名称反向澄清用户仅提供专利名称、标题或一句话方向专利名称反向澄清卡 + 人工确认记录
L1 代码证据映射已有候选方案，需判断是否落在代码实现上方案-代码证据映射表
L2 技术交底书当前最推荐主产物算法/软件类说明书式技术交底书 + 必要证据附录
L2.5 权利要求布局需要推进到申请文件层但先稳住 claim tree权利要求布局卡 + 权利要求-证据矩阵
L3 发明专利初稿需要代理撰写前的完整初稿说明书初稿 + 权利要求草稿 + 摘要 + 自检表
L4 可专利方案挖掘用户尚未总结候选方案候选方案清单 + 优先级建议 + 人工筛选记录
LevelApplicable ScenarioMain Deliverables
L0 Title Reverse ClarificationUser only provides patent name, title, or one-sentence directionPatent Title Reverse Clarification Card + Manual Confirmation Record
L1 Code Evidence MappingCandidate solutions exist, need to judge whether they fall within code implementationSolution-Code Evidence Mapping Table
L2 Technical Disclosure DocumentCurrently the most recommended main deliverableAlgorithm/Software Specification-Style Technical Disclosure Document + Necessary Evidence Appendices
L2.5 Claim LayoutNeed to advance to the application document layer but stabilize the claim tree firstClaim Layout Card + Claim-Evidence Matrix
L3 Invention Patent DraftComplete draft required before agent writingSpecification Draft + Claim Draft + Abstract + Self-Check List
L4 Patentable Solution MiningUser has not summarized candidate solutions yetCandidate Solution List + Priority Recommendations + Manual Screening Record

标准执行顺序

Standard Execution Sequence

按目标层级裁剪执行，不必每次输出全部文件。
  1. 预读用户材料，确认目标层级和模板要求。
  2. 创建 
    archive/YYYY-MM-DD-主题/
    。
  3. 判断输入成熟度：T0/T1 先澄清，无清单先挖掘。
  4. 读取项目 README、PRD、架构和依赖文件，形成边界画像。
  5. 围绕已确认方案读取代码，提取证据并分级。
  6. 将工程实现转译为 S1...Sn、技术特征、技术效果和替代实施方式。
  7. 生成交底书，正文采用说明书式结构，证据后置。
  8. 如需申请文件层，生成布局卡和证据矩阵。
  9. 如需完整初稿，生成发明专利初稿和自检表。
  10. 输出待研发、代理师或申请主体确认的问题。
Tailor execution according to the target level, there is no need to output all files each time.
  1. Pre-read user materials to confirm the target level and template requirements.
  2. Create 
    archive/YYYY-MM-DD-Theme/
    .
  3. Judge input maturity: clarify first for T0/T1, mine first if there is no list.
  4. Read project README, PRD, architecture, and dependency files to form a boundary profile.
  5. Read code around confirmed solutions, extract evidence and grade it.
  6. Translate engineering implementations into S1...Sn, technical features, technical effects, and alternative implementations.
  7. Generate the disclosure document, with the body adopting a specification-style structure and evidence placed at the end.
  8. If application document layer is required, generate layout card and evidence matrix.
  9. If a complete draft is required, generate the invention patent draft and self-check list.
  10. Output questions to be confirmed by R&D, patent attorneys, or application entities.

推荐交付文件

Recommended Delivery Files

text
archive/YYYY-MM-DD-主题/
├── 00-输入材料摘要.md
├── 01-专利名称反向澄清卡-待确认.md
├── 01A-人工确认记录.md
├── 02-项目边界与依赖画像.md
├── 03-项目技术方案画像.md
├── 04-方案代码证据映射-<方案名>.md
├── 05-技术交底书-<方案名>.md
├── 05A-权利要求布局卡-<方案名>.md
├── 05B-权利要求-证据矩阵-<方案名>.md
├── 06-发明专利初稿-<方案名>.md
├── 06A-发明专利初稿自检表-<方案名>.md
└── 07-研发补充问题清单.md
若当前是候选方案挖掘模式，先输出 
03-候选可专利方案清单-待人工筛选.md
和 
04-优先申请建议.md
，待用户确认后再进入定向检索。
text
archive/YYYY-MM-DD-Theme/
├── 00-Input Materials Summary.md
├── 01-Patent Title Reverse Clarification Card-Pending Confirmation.md
├── 01A-Manual Confirmation Record.md
├── 02-Project Boundary and Dependency Profile.md
├── 03-Project Technical Solution Profile.md
├── 04-Solution Code Evidence Mapping-<Solution Name>.md
├── 05-Technical Disclosure Document-<Solution Name>.md
├── 05A-Claim Layout Card-<Solution Name>.md
├── 05B-Claim-Evidence Matrix-<Solution Name>.md
├── 06-Invention Patent Draft-<Solution Name>.md
├── 06A-Invention Patent Draft Self-Check List-<Solution Name>.md
└── 07-R&D Supplementary Question List.md
If in candidate solution mining mode, first output 
03-Candidate Patentable Solution List-Pending Manual Screening.md
and 
04-Priority Application Recommendations.md
, then enter targeted retrieval after user confirmation.

交底书默认结构

Default Structure of Disclosure Document

软件、算法和 Agent 系统方案的技术交底书默认使用以下顺序：
  1. 方案基本信息
  2. 技术领域
  3. 背景技术
  4. 发明内容
  5. 附图说明
  6. 具体实施方式
  7. 技术效果
  8. 替代实施方式
  9. 代码证据摘要与附录索引
  10. 待研发 / 代理师确认事项
如果正文中连续出现多个代码路径、函数名或字段名，应回退重写，将其移入 
04-方案代码证据映射-<方案名>.md
或证据附录。
The technical disclosure document for software, algorithm, and Agent system solutions uses the following order by default:
  1. Basic Solution Information
  2. Technical Field
  3. Background Technology
  4. Invention Content
  5. Description of Drawings
  6. Specific Implementation Methods
  7. Technical Effects
  8. Alternative Implementation Methods
  9. Code Evidence Summary and Appendix Index
  10. Items to be Confirmed by R&D / Patent Attorneys
If multiple code paths, function names, or field names appear consecutively in the body, rewrite them by moving them to 
04-Solution Code Evidence Mapping-<Solution Name>.md
or evidence appendices.

输入/输出

Input/Output

输入

Input

  • 必需：代码文件或项目目录、技术描述、创新点说明、候选专利方案，至少具备其中两类；若只有专利名称，按 L0 处理。
  • 可选：PRD、需求说明、会议纪要、技术领域、应用场景、附图偏好、输出层级、申请主体信息、发明人候选、优先权信息、客户模板。
  • Required: At least two of code files or project directories, technical descriptions, innovation point explanations, candidate patent solutions; if only patent names are provided, handle according to L0.
  • Optional: PRD, requirement description, meeting minutes, technical field, application scenarios, drawing preferences, output level, application entity information, candidate inventors, priority information, customer templates.

输出

Output

  • 专利名称反向澄清卡
  • 候选可专利方案清单
  • 项目边界与依赖画像
  • 方案代码证据映射表
  • 算法/软件类说明书式技术交底书
  • 权利要求布局卡
  • 权利要求-证据矩阵
  • 发明专利初稿
  • 发明专利初稿自检表
  • 待确认事项清单
  • Patent Title Reverse Clarification Card
  • List of Candidate Patentable Solutions
  • Project Boundary and Dependency Profile
  • Solution Code Evidence Mapping Table
  • Algorithm/Software Specification-Style Technical Disclosure Document
  • Claim Layout Card
  • Claim-Evidence Matrix
  • Invention Patent Draft
  • Invention Patent Draft Self-Check List
  • List of Pending Confirmation Items