agentsclimarketplace

Academic nature nature paper to patent

Skill hamzabellouch/agent-skills/Academic and Scientific Research/academic-nature-nature-paper-to-patent

Convert scientific papers, theses, technical reports, source code, figures, inventor notes, or research manuscripts into evidence-grounded Chinese invention patent drafts and attorney-facing technical disclosure materials. Use when an AI agent must mine patent points, draft or revise a Chinese technical disclosure (技术交底书), run prior-art comparison, convert Office project materials, map every claimed feature to source evidence, preserve core formulas as editable Office Math, generate claim-aligned flowcharts and methodology figures, compare a paper with an existing patent, audit support and consistency, or deliver Chinese DOCX patent/disclosure files.From its SKILL.md

Install
npx -y skills add hamzabellouch/agent-skills --skill academic-nature-nature-paper-to-patent

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 3 stars3 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

4.9 KB, 916 tokens by cl100k_base, as published. Nobody here has run it

Paper to Chinese Patent

Use this file as the router for the patent-drafting workflow. Do not draft the application directly from the paper abstract or contribution list.

1. Load the workflow

Read manifest.yaml, then read every file under always_load.

Detect these axes from the user's files and request:

  • source_format: selectable PDF, scanned PDF, pasted text, or mixed project;
  • task_mode: full draft, claim set, disclosure analysis, technical disclosure, disclosure iteration, or paper-patent audit;
  • invention_type: algorithm/software, apparatus/system, process/material, or mixed.

State the detected values in one short line. Load only the matching fragments declared in the manifest. Load detailed references only when their condition applies.

2. Preserve source grounding

Create stable source IDs before drafting:

  • P001... for paper text blocks;
  • E001... for equations;
  • F001... for source figures;
  • C001... for source-code or supplementary evidence.

Every material feature in a formal claim must map to one or more source IDs. Use only explicit, inherent, needs-confirmation, or unsupported as support states. Exclude unsupported features from formal claims.

Never infer inventorship, ownership, unpublished implementation details, publication dates, prior-art conclusions, or legal sufficiency. Use [TO CONFIRM: specific question] outside formal claims when facts are missing.

3. Draft through stage gates

For full-draft, claim-set, disclosure-analysis, and paper-patent-audit, complete the stages in static/core/workflow.md in order. Persist the intermediate artifacts specified there. Do not move to formal claims until the source map, terminology ledger, inventories, evidence ledger, and invention concept pass their gates.

For technical-disclosure, follow the ordered prompt references in static/fragments/task/technical-disclosure.md. For disclosure-iteration, follow static/fragments/task/disclosure-iteration.md and preserve the prior draft instead of restarting the formal application workflow.

For a full application, draft claims first, then align the specification, figures, embodiments, and abstract to the claim terminology and step order.

4. Produce Chinese formal documents

Agent-facing analysis may use the user's preferred language. Produce formal Chinese patent deliverables in Chinese when the task is a formal application package:

  • 权利要求书;
  • 说明书;
  • 说明书摘要;
  • 摘要附图;
  • figure labels and descriptions.

For technical-disclosure and disclosure-iteration, produce the Chinese technical disclosure (技术交底书) as timestamped Markdown plus matching DOCX, with Mermaid system/process diagrams rendered through scripts/disclosure/.

For algorithmic inventions, retain source-supported core formulas, define every symbol, explain each formula's technical operation, and render formulas as native editable Office Math in DOCX. Do not use plain LaTeX strings as the visible formula.

Generate the main flowchart from the ordered steps of the principal method claim. Its final node must name the concrete domain output, such as a defect detection result, target pose, state estimate, or control instruction. Reuse the same main figure as the abstract figure and a specification figure.

5. Validate before delivery

For formal application packages, populate the structured draft described in references/draft-schema.md, then run:

python scripts/validate_patent_draft.py draft.json
python scripts/build_patent_package.py draft.json --output-dir outputs --prefix patent

Resolve all validation ERROR findings. Review every WARNING against the source. Label the result incomplete draft when a required quality threshold in static/core/output-contract.md is not met.

For technical disclosures, run the internal checks in references/disclosure/disclosure_self_check.md, render Mermaid/Word outputs with scripts/disclosure/mermaid_render.py, and resolve formula, parameter, prior-art URL, and chapter-consistency issues before delivery.

The generated package is a drafting aid for inventor and patent-professional review, not a patentability opinion, infringement opinion, or filing guarantee.

What ships with it: 60 files

283.7 KB alongside SKILL.md, 18 of them executable

evals/

20 more files not listed here. See all 60 in the repository.

Gives 0 of the 12 instructions most legal skills give in 916 tokens

Counted across 234 of the 234 authors here whose files we hold, read 2026-08-07

  • Use text operators for text fieldsin 11 of 234, across 6 files
  • Consult qualified counsel before usein 11 of 234, across 3 files
  • Use PatentSearch API for patent searchesin 10 of 234, across 5 files
  • Confirm jurisdiction, employment type, and required clausesin 9 of 234, across 2 files
  • Choose a document template and tailor role-specific termsin 9 of 234, across 2 files
  • Validate compensation, benefits, and compliance requirementsin 9 of 234, across 2 files
  • Add signature, confidentiality, and IP assignment terms as neededin 9 of 234, across 2 files
  • Open the implementation playbook for detailed templatesin 9 of 234, across 2 files
  • Use TSDR for trademark data retrievalin 9 of 234, across 4 files
  • Ask for clarification if required inputs are missingin 8 of 234, across 2 files
  • Set the USPTO_API_KEY environment variablein 8 of 234, across 3 files
  • Use the uspto-opendata-python library for PEDSin 8 of 234, across 3 files

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.