patent-novelty-check

Q: How do I install patent-novelty-check?

Run npx skills add https://github.com/wanshuiyin/auto-claude-code-research-in-sleep --skill patent-novelty-check in your project. The skills CLI detects the agents you use and copies the skill into the right folder. For Claude Code you can also place the skill folder at ~/.claude/skills/patent-novelty-check/ (personal) or .claude/skills/patent-novelty-check/ (project).

by wanshuiyinMedia & Creative

Assess patent novelty and non-obviousness against prior art. Use when user says "专利查新", "patent novelty", "可专利性评估", "patentability check", or wants to evaluate if an invention is patentable.

Installs: 509
Last 24h: —
Trending rank: —
Security: Not audited

Install

Works with every agent the skills CLI supports.

npx skills add https://github.com/wanshuiyin/auto-claude-code-research-in-sleep --skill patent-novelty-check

01
Install
Run the command in your project. The CLI asks which agents to install for.
02
Describe the task
Ask your agent for something this skill covers. It loads patent-novelty-check automatically when relevant.
03
Review the result
The agent follows the SKILL.md steps, scripts and references bundled with the skill.

SKILL.md

The instructions your agent reads when the skill is triggered.

Patent Novelty and Non-Obviousness Check

Before each reviewer slot, follow Selectable Reviewers: resolve invocation → .aris/reviewer.json → Claude Code default (Codex MCP) with reviewer_route.py through the canonical helper resolver, preserving this slot's tier, fresh contexts, source/image inputs and deterministic gates. Codex examples/pins below apply only to backend=codex; other routes use their own first-call/reply contract. Save the route and actual returned identity; unavailable selections are REVIEW_UNAVAILABLE. Classify acceptance from actual families/completion/scope. Forward invocation reviewer/model/effort overrides to reviewer-bearing children. Generation workers are unchanged.

Assess patentability of: $ARGUMENTS

Adapted from /novelty-check for patent legal standards. Research novelty is NOT the same as patent novelty.

Constants

REVIEWER_MODEL = gpt-6-astra — Model used via Codex MCP for cross-model examiner verification
NOVELTY_STANDARD = patent — Always use legal patentability standard, not research contribution standard

Inputs

Invention description from $ARGUMENTS
patent/PRIOR_ART_REPORT.md (output of /prior-art-search)
patent/INVENTION_BRIEF.md if exists

Shared References

Load ../shared-references/patent-writing-principles.md for novelty/non-obviousness standards. Load ../shared-references/patent-format-us.md for 102/103 analysis framework.

Workflow

Step 1: Define Claim Elements

From the invention description, extract the key claim elements that would define the invention's scope:

List the technical features that make the invention novel
Identify which features are known from prior art vs. inventive
Draft preliminary claim language for 2-3 independent claims (method + system)

Step 2: Anticipation Analysis (Novelty)

For each preliminary claim, test against EACH prior art reference in PRIOR_ART_REPORT.md:

Single-reference test: Does any single reference disclose ALL claim elements?

Claim Element	Ref 1	Ref 2	Ref 3	...
Feature A	Yes/No + evidence
Feature B	Yes/No + evidence
Feature C	Yes/No + evidence
Feature D	Yes/No + evidence

Verdict per reference:

ANTICIPATED: One reference discloses every element → claim is not novel
NOT ANTICIPATED: At least one element missing from every single reference → claim is novel

Step 3: Obviousness Analysis (Inventive Step)

If the invention is novel (passes Step 2), test for obviousness:

Two/three-reference combination test: Can 2-3 references be combined to render the claim obvious?

For each combination of the top references:

Primary reference: Which reference is closest to the claimed invention?
Secondary reference(s): Which reference(s) teach the missing element(s)?
Motivation to combine: Would a POSITA have reason to combine these references?
- Explicit suggestion in the references themselves?
- Same field, same problem?
- Common design incentive?
- Known technique for improving similar devices?

Format as a matrix:

Combination	Primary	Secondary	Missing Elements	Motivation to Combine	Obvious?
Ref1 + Ref2	Ref1	Ref2	Feature D	Same field, similar problem	Yes/No

Step 4: Cross-Model Examiner Verification

Call REVIEWER_MODEL via mcp__codex__codex with xhigh reasoning:

Resolve this reviewer slot first; the following is the backend=codex example.

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

    INVENTION: [invention description + preliminary claims]

    PRIOR ART: [prior art references with key teachings]

    Please analyze:
    1. Anticipation (novelty): Does any single reference anticipate any claim?
    2. Obviousness: Can any combination of references render claims obvious?
    3. Claim scope: Are the claims broad enough to be valuable?
    4. Recommended amendments if any claim is rejected.
    Be rigorous and cite specific references.

Step 5: Jurisdiction-Specific Assessment

For each target jurisdiction, provide a patentability assessment:

Under 35 USC 102/103 (US):

Novelty: PASS / FAIL (cite specific reference if fail)
Non-obviousness: PASS / FAIL (cite combination if fail)

Under Article 22 CN Patent Law (CN):

新颖性 (Novelty): 通过 / 未通过
创造性 (Inventive Step): 通过 / 未通过

Under Article 54/56 EPC (EP):

Novelty: PASS / FAIL
Inventive step: PASS / FAIL (problem-solution approach)

Step 6: Output

Write patent/NOVELTY_ASSESSMENT.md:

## Patentability Assessment

### Invention Summary
[description]

### Overall Assessment
[PATENTABLE / PATENTABLE WITH AMENDMENTS / NOT PATENTABLE]

### Anticipation Analysis
[claim-by-claim matrix against each reference]

### Obviousness Analysis
[combination analysis with motivation to combine]

### Cross-Model Examiner Review
[summary of GPT-6-Astra examiner feedback]

### Recommended Claim Amendments
[If claims need modification to overcome prior art, suggest specific amendments]

### Risk Factors
[What could cause rejection during actual prosecution?]

Key Rules

Patent novelty is absolute: any public disclosure before the priority date counts as prior art, worldwide.
Research novelty ("has anyone published this?") is NOT the same as patent novelty ("does any single reference teach every claim element?").
Obviousness requires BOTH: (1) a combination of references AND (2) a motivation to combine them.
Never assume the invention is patentable just because no identical patent exists.
The assessment is advisory only -- actual prosecution may reveal different prior art.
If the selected reviewer is unavailable, record REVIEW_UNAVAILABLE; continue only as an explicitly unreviewed draft, without switching providers.

FAQ

What does the patent-novelty-check skill do?

How do I install patent-novelty-check?

Run npx skills add https://github.com/wanshuiyin/auto-claude-code-research-in-sleep --skill patent-novelty-check in your project. The skills CLI detects the agents you use and copies the skill into the right folder. For Claude Code you can also place the skill folder at ~/.claude/skills/patent-novelty-check/ (personal) or .claude/skills/patent-novelty-check/ (project).

Which AI agents can use patent-novelty-check?

Any agent that supports the open Agent Skills format (a folder with a SKILL.md file), including Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, Windsurf, Cline, OpenCode and more.

Is patent-novelty-check safe to install?

No partner security audit is available yet. Skills can run commands and scripts, so read the SKILL.md and any bundled files before installing.

Who maintains patent-novelty-check?

It is published by wanshuiyin in wanshuiyin/auto-claude-code-research-in-sleep. It has 509 recorded installs.