Claims Drafting logo

Claims Drafting

CommunityPopular
wanshuiyin
claims-drafting

Draft patent claims for an invention. Use when user says "撰写权利要求", "draft claims", "写权利要求书", "claim drafting", or wants to create patent claims. The core skill of the patent pipeline.

Overview

Publisherwanshuiyin
RepositoryAuto-claude-code-research-in-sleep
Skill nameclaims-drafting
Stars
16.3K
Forks
1.4K
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by wanshuiyin on GitHub. Read the source before you install it.

Installation

Install the Claims Drafting AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/wanshuiyin/Auto-claude-code-research-in-sleep.git /tmp/Auto-claude-code-research-in-sleep
mkdir -p .claude/skills
cp -r /tmp/Auto-claude-code-research-in-sleep/skills/claims-drafting .claude/skills/claims-drafting
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Claims Drafting in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Claims Drafting on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Claims Drafting is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Claims Drafting: The Core Patent Skill

Draft patent claims based on: $ARGUMENTS

This is the most critical skill in the patent pipeline. Claims define the legal scope of protection -- everything else (specification, figures, abstract) exists to support and enable the claims.

Constants

  • REVIEWER_MODEL = gpt-6-astra — External examiner for claim quality review
  • MAX_CLAIM_REVISION_ROUNDS = 3 — Maximum revision iterations
  • CLAIM_STYLE = "auto"US (Jepson or open), EP (two-part mandatory), CN (two-part), auto (detect from jurisdiction)
  • MIN_INDEPENDENT_CLAIMS = 2 — Typically method + system. For utility model (实用新型): apparatus/device only, NO method claims.
  • MAX_TOTAL_CLAIMS = 20 — Practical limit (USPTO includes 20 in base fee)
  • PATENT_TYPE = "invention"invention (发明专利) or utility_model (实用新型, apparatus claims only)

Inputs

  1. patent/INVENTION_DISCLOSURE.md — structured invention with core/supporting/optional features
  2. patent/PRIOR_ART_REPORT.md — prior art to avoid
  3. patent/NOVELTY_ASSESSMENT.md — novelty analysis with suggested amendments
  4. Target jurisdiction from invention disclosure or $ARGUMENTS

Shared References

Load ../shared-references/patent-writing-principles.md for claim drafting principles, antecedent basis rules, and common pitfalls. Load ../shared-references/patent-format-cn.md for CN claim format (其特征在于). Load ../shared-references/patent-format-us.md for US claim format (comprising, means-plus-function). Load ../shared-references/patent-format-ep.md for EP two-part form (characterised in that).

Workflow

Step 1: Determine Claim Style and Patent Type

Based on patent type and jurisdiction:

If PATENT_TYPE = utility_model (实用新型):

  • CN jurisdiction ONLY
  • Apparatus/device claims ONLY — no method, no product-by-process
  • MIN_INDEPENDENT_CLAIMS = 1 (single apparatus claim is sufficient)
  • Claim format: "1. 一种[主题],其特征在于,包括:[组件描述]。"

Based on target jurisdiction:

JurisdictionClaim StyleCharacterising PhrasePreamble Format
CNTwo-part (两部式)其特征在于一种...的方法/装置,包括:
USOpen (preferred)comprisingA method for..., comprising:
EPTwo-part (mandatory)characterised in thatA method for..., comprising [known], characterised in that [inventive]
ALLDraft CN + US + EPAll of the aboveAll of the above

Step 2: Draft Independent Claims

CRITICAL — Claims numbering (CN format):

  • Claims must be numbered 1, 2, 3, ... continuously without gaps
  • Independent and dependent claims are INTERMIXED in final numbering
  • Do NOT group independent claims separately from dependent claims
  • Example correct: Claim 1 (independent, product), Claim 2 (depends on 1), Claim 3 (depends on 1), Claim 4 (independent, method), Claim 5 (depends on 4)...
  • Example WRONG: Claim 1 (independent), Claim 5 (independent), Claim 8 (independent), then Claim 2, 3, 4, 6, 7, 9, 10 as dependents

CRITICAL — No empirical content in claims:

  • Claims describe ONLY structural features or method steps
  • Do NOT include signal characteristics, detection principles, measurement results
  • WRONG: "产生负脉冲信号" / "谐振频率下降" — these are results, not features
  • RIGHT: "所述开口谐振环的开口处形成用于检测通过流体中颗粒物的间隙传感区域"

For each claim category identified in INVENTION_DISCLOSURE.md:

Method Claim (broadest):

  1. Start with preamble identifying the category and purpose
  2. List the core inventive features (from the "Core Inventive Concept" section)
  3. Include enough known features for context (but not more than necessary)
  4. Use open transition ("comprising" / "包括")
  5. Each element should be separated by semicolons or on separate lines
  6. Apply jurisdiction-specific format:
    • CN: 前序部分 + "其特征在于" + 特征部分
    • US: Preamble + "comprising:" + elements
    • EP: Preamble + known features + "characterised in that" + inventive features

System/Apparatus Claim:

  1. Mirror the method claim in structural form
  2. Each method step becomes a "module configured to..." or "component for..."
  3. Same hierarchy of known vs. inventive features

Quality checks for each independent claim:

  • Single sentence (US/EP) or properly structured (CN)
  • Antecedent basis: "a" first, "the" thereafter for each element
  • No relative terms without definition
  • No result-to-be-achieved limitations
  • Transitional phrase is appropriate (open preferred)
  • Preamble does not import unnecessary limitations
  • Each element is necessary for patentability
  • Claim scope is broadest defensible over prior art

Step 3: Draft Dependent Claims

For each independent claim, draft 5-10 dependent claims that:

  1. Narrow the core inventive features: Specific implementations, parameters, ranges
  2. Cover preferred embodiments: Features from the specification's detailed description
  3. Provide fallback positions: If the independent claim is rejected, these narrower claims may survive
  4. Cover alternatives: Different ways to achieve the same inventive result

Dependent claim format:

  • CN: "根据权利要求X所述的[主题],其特征在于,所述[特征]具体为..."
  • US: "The [method/system] of claim X, wherein the [element] comprises [limitation]."
  • EP: "The [method/system] according to claim X, characterised in that [limitation]."

Rules for dependent claims:

  • Each must add at least one meaningful limitation
  • Must reference a prior claim by number
  • Must not merely repeat the parent claim
  • Should not be cumulative (each claim should be independently useful as a fallback)
  • Multiple dependent claims (US: only "or" references, not "and")

Step 4: Claim-to-Specification Mapping

Create a preliminary mapping to verify enablement:

Claim ElementMust be described in specificationReference numeral
[element 1]Yes/No[numeral]
[element 2]Yes/No[numeral]

If any element lacks specification support, add it to the specification requirements.

Step 5: Cross-Model Examiner Review

Call REVIEWER_MODEL via mcp__codex__codex with xhigh reasoning:

mcp__codex__codex:
  model: gpt-6-astra
  config: {"model_reasoning_effort": "xhigh"}
  prompt: |
    You are a senior patent examiner at the [USPTO/CNIPA/EPO].
    Review the following patent claims for quality and patentability.

    CLAIMS: [all claims]

    PRIOR ART: [prior art references from PRIOR_ART_REPORT.md]

    INVENTION: [summary from INVENTION_DISCLOSURE.md]

    Analyze each claim for:
    1. Clarity (35 USC 112(b) / Art 84 EPC): Are terms definite?
    2. Written description support: Does the spec support all claim scope?
    3. Anticipation (102/Art 54): Would any single reference anticipate?
    4. Obviousness (103/Art 56): Would any combination render obvious?
    5. Claim scope: Are independent claims broad enough to be valuable?
    6. Dependent claims: Do they provide meaningful fallback positions?
    7. Antecedent basis: Any issues with "a"/"the" usage?
    8. Indefinite terms: Any functional/result language issues?

    For each issue found, provide:
    - The specific claim number and element
    - The problem (cite statute/rule)
    - A suggested fix

    Provide an overall PATENTABILITY SCORE: 1-10.

Step 6: Revision Loop

If the examiner review identifies issues:

  1. Address all CRITICAL issues (anticipation, obviousness, indefiniteness)
  2. Address MAJOR issues (scope too narrow, missing support, weak fallbacks)
  3. Consider MINOR issues (antecedent basis, formatting)
  4. Re-submit to examiner for round 2 (use mcp__codex__codex with threadId)
  5. Repeat up to MAX_CLAIM_REVISION_ROUNDS times

Step 7: Output

Write patent/CLAIMS.md:

markdown
## Patent Claims

### Independent Claims

#### Claim 1 — Method
[formatted claim text]

#### Claim X — System/Apparatus
[formatted claim text]

### Dependent Claims

#### Claim 2 (depends on 1)
[formatted claim text]

#### Claim 3 (depends on 1)
[formatted claim text]
...

### Claims Summary Table

| Claim | Type | Depends On | Key Limitation | Prior Art Avoidance |
|-------|------|-----------|----------------|---------------------|
| 1 | Method || [core inventive features] | [what makes it novel over prior art] |
| 2 | Method | 1 | [narrowing] | [additional distinguishing] |
| X | System || [mirrors claim 1] | [same as claim 1] |
...

### Examiner Review Summary
[Key findings and how they were addressed]

Key Rules

  • Claims are the single most important part of the patent. Everything else supports them.
  • Draft independent claims first, then dependent claims.
  • Independent claims must be broadest defensible scope over prior art -- not broader, not narrower.
  • Each dependent claim should be independently useful as a fallback position.
  • Antecedent basis is mandatory: "a processor" first, "the processor" thereafter.
  • Use "comprising" (open) unless there is a specific reason for "consisting of" (closed).
  • Never include result-to-be-achieved language in claims ("configured to achieve high accuracy").
  • Never fabricate claim language -- every element must come from the actual invention.
  • If drafting for ALL jurisdictions, produce separate claim sets for CN, US, and EP.
  • If mcp__codex__codex is not available, skip cross-model examiner review and note it in the output.

Frequently asked questions

What does the Claims Drafting AI skill do?

Draft patent claims for an invention. Use when user says "撰写权利要求", "draft claims", "写权利要求书", "claim drafting", or wants to create patent claims. The core skill of the patent pipeline.

Why use Claims Drafting on TypingMind?

Because you install it once and use it with any model. Claims Drafting is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Claims Drafting in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/wanshuiyin/Auto-claude-code-research-in-sleep/tree/main/skills/claims-drafting. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Claims Drafting?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Claims Drafting?

As many as you like. As long as a model supports skills, you can use Claims Drafting with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Claims Drafting AI skill free?

Yes. It is published on GitHub by wanshuiyin under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇