| name | research-nature-writing |
| description | Use when planning, drafting, restructuring, or polishing high-impact journal style academic prose, Nature-style manuscript sections, article abstracts, introductions, discussions, contribution statements, or publication-leaning research writing. |
Research Nature Writing
This skill adapts high-impact article-writing logic into a safer general research-writing layer.
It improves argument presentation. It does not create evidence.
Dependency Chain
For formal, supervisor/PI/client/reviewer-facing, or submission-facing writing, use this skill only after:
- source-first checks;
- compliance/privacy checks where relevant;
- cognitive-framework planning;
- source-readiness and citation checks where relevant;
- academic self-review loop for formal prose.
Then use:
- relevant academic/professional style checks;
- document-quality and delivery gates.
Existing Skill Boundary
| Existing Skill | Boundary |
|---|
uk-academic-writing-style | Controls British academic tone when relevant |
academic-self-review-loop | Judges intrinsic paragraph quality before style polish |
dissertation-literature-review | Controls literature synthesis and source clusters |
dissertation-citation-audit | Checks citations and claim support |
research-nature-writing | Adds high-impact article logic: sharper problem framing, contribution line, paragraph movement, and reader journey |
Workflow
- Identify the writing object:
- abstract;
- introduction;
- literature review / related work;
- methods;
- discussion;
- conclusion;
- contribution statement;
- title or highlights.
- State the claim, evidence status, and audience.
- Improve structure before wording:
- problem;
- gap or bottleneck;
- move made by this study;
- evidence or method warrant;
- contribution boundary.
- Improve paragraph flow:
- one controlling point per paragraph;
- topic sentence as a mini-claim;
- citations serve the point, not a reading list;
- final sentence moves the argument forward.
- Polish wording only after evidence and logic are stable.
Guardrails
- Do not invent citations, results, methods, participants, datasets, reviewer feedback, or institutional requirements.
- Do not promise publication readiness, acceptance, official marks, funding, client approval, or AI-detector outcomes.
- Do not import hype, overclaiming, or novelty inflation from journal style.
- Preserve project-specific caution and compliance requirements.
- If source support is weak, weaken the claim rather than strengthening the language.