
Nature Paper Card
Turns one scientific paper into a structured, evidence-linked deep-reading card with method and limitation analysis
- Repository
- Yuan1z0825/nature-skills
- Stars
- 51,827
- Licence
- Apache-2.0
- Files
- 15 files
What it does
When installed, the agent produces a structured 16-section deep-reading card for a single scientific paper. Instead of a plain summary, it breaks down the methods, links each central claim to the experiments or figures that support it, lists author-stated and agent-identified limitations, and proposes research ideas framed as hypotheses to test. The skill runs bundled scripts to prepare the paper text and to audit the finished card against the source for groundedness.
Great for
- Researchers who need to critically read a paper and check whether its claims are backed by its experiments
- Graduate students preparing literature notes or journal-club material
- Scientists scouting for follow-up research questions from a single paper
Skip it if
- Full bilingual translation of a paper, since the skill explicitly excludes it (the description says so)
- Formal peer review or publication-ready articles, which the skill is scoped out of
- Papers supplied only as abstracts or metadata, where many sections will be marked not assessable
Try asking
- “Make a deep-reading Paper Card for this PDF, focusing on the method and experiment-to-claim evidence”
- “Analyse the limitations of this paper and whether its main claims are supported by the ablations”
- “Build a 论文精读卡 for the attached paper and list possible research ideas as hypotheses”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 15 original files, the Apache-2.0 licence and an attribution note.
nature-paper-card.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip nature-paper-card.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip nature-paper-card.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Supply the full PDF with figures and tables rather than the abstract alone, so the evidence inventory is complete
- Check the audit report warnings yourself; the skill asks the agent to review warnings with judgment rather than suppress them
- Ask for the card in your language; the skill matches the user's language by default
What's inside
- agents/openai.yaml
- evals/evals.json
- manifest.yaml
- README_EN.md
- README.md
- references/card-schema.md
- references/evidence-and-provenance.md
- references/research-idea-gates.md
- scripts/audit_paper_card.py
- scripts/prepare_paper.py
- SKILL.md
- static/core/output-contract.md
- static/core/principles.md
- static/core/workflow.md
- static/fragments/paper_type/clinical.md
Redistributed unmodified from Yuan1z0825/nature-skills under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.