# Draft Nature-style manuscript sections from your notes

Researchers often have raw claims, figures, or Chinese drafts but struggle to produce well-structured Nature-style prose. This skill loads the right writing fragments and drafts rigorous, publication-ready manuscript sections.

## Install

```bash
npx skillstore add community contribution, refactored into static/dynamic layers/yuan1z0825-nature-writing
```

## Metadata

- - Slug: yuan1z0825-nature-writing
- - Version: 1.0.0
- - Author: Community contribution, refactored into static/dynamic layers
- - GitHub username: Yuan1z0825
- - License: MIT
- - Repository: https://github.com/Yuan1z0825/nature-skills/tree/main/skills/nature-writing
- - Ref: main
- - Supported tools: Claude, Codex, Claude Code
- - Risk level: low
- - Risk factors: filesystem
- - Quality score: 79
- - Quality tier: bronze
- - Public page: https://skillstore.pages.dev/skills/yuan1z0825-nature-writing
- - Manifest: https://skillstore.pages.dev/api/skills/yuan1z0825-nature-writing/manifest

## Capabilities

- Drafts Nature-style abstract, introduction, method, experiments, discussion, conclusion, and title sections
- Plans the one-sentence argument and argument chain before writing prose
- Loads paper-type-specific playbooks for research, methods, hypothesis, algorithmic, and review papers
- Supports both English notes and Chinese-source notes with zh-to-en translation intent
- Adjusts framing for Nature, Nature Communications, or generic scientific journals
- Self-audits drafts against an adversarial reviewer checklist for rejection risks

## Use Cases

- Draft a full manuscript from lab notes: A graduate student has figures, results tables, and Chinese notes. The skill detects axes, loads matching fragments, and drafts all required sections in Nature style.
- Rebuild the introduction for resubmission: Reviewers asked for a sharper contribution framing. The skill rewrites the introduction using a challenge-decomposition pipeline and a Nature summary-paragraph funnel.
- Generate a methods section from a technical outline: An engineer has a pipeline diagram and bullet points. The skill organizes them into a module-by-module methods section with motivation, design, and advantages.

## Prompt Templates

### Draft an abstract from claims and results

```
I have the following claims and key results from my experiments. Draft a Nature-style abstract following the background-gap-method-result-significance-boundary structure.

Claims: [paste 3-5 sentences]
Results: [paste key numbers, figures, or ablations]
Journal target: Nature Communications
```

### Restructure an introduction using a challenge funnel

```
My current introduction reads like a literature review. Restructure it using the challenge-decomposition pipeline: field-scale gap, bottleneck of existing work, unresolved problem, our contribution. Here is the draft:

[paste introduction]
```

### Write a methods section from a module outline

```
Generate a Nature-style methods section from this pipeline outline. For each module explain motivation, design, forward flow, and technical advantages.

Modules: [list modules]
Figures referenced: [list figure numbers]
```

### Run a rejection-risk self-review

```
Audit my manuscript against an adversarial reviewer checklist. Identify weak claims, unsupported mechanisms, missing ablations, and ambiguous boundary statements.

[paste full draft or specific section]
```

## Limitations

- Does not polish sentence-level English; for that, use a dedicated polishing skill
- Cannot invent data, statistics, mechanisms, or novelty that the author has not provided
- Does not perform literature search or generate citations automatically
- Cannot replace domain expertise; final claims must be verified by the author

## Best Practices

- State the detected axes in one short line before drafting so the user can correct cheaply
- Write the one-sentence argument first, then build sections around it
- Load only the fragments the axes select, never the full static layer
- List missing evidence under 'Assumptions or missing inputs' instead of inventing content
- Run the adversarial self-review before considering a draft complete

## Anti Patterns

- Do not invent data, mechanisms, statistics, sample sizes, or novelty claims
- Do not load every fragment in static/ into context for a single request
- Do not write a Chinese-source draft by mirroring the Chinese sentence order into English
- Do not skip the planning steps even when the user asks for prose immediately

## Security Audit

- - Safe to publish: true
- - Audited at: 2026-06-24T06:10:15.552\+00:00
- - Summary: The static scanner flagged 776 potential issues across 66 files, but manual review confirms nearly all findings are false positives. The skill is an academic writing assistant that drafts manuscript sections. The 'weak cryptographic algorithm' hits refer to academic vocabulary like 'hash', 'seed', and 'digest' used in writing contexts. The 'path traversal' hits are YAML fragment-path references in the manifest. The 'Ruby/shell backtick execution' hits are markdown inline-code formatting throughout prose and examples. The only real filesystem behavior is reading skill fragment files. No prompt injection attempts, no exfiltration, and no malicious intent were detected.

## Stats

- - Views: 0
- - Downloads: 13
- - Favorites: 0
- - Popularity score: 0