| name | paper-writing |
| description | Write a scientific manuscript from research notes and experimental results. Use when drafting, revising, or structuring an academic paper, report, or research summary. |
Paper Writing
Compose a scientific manuscript that reflects what was actually found — no more, no less.
Inputs
Require: the research question, the literature synthesis (from literature-review), the hypotheses (from hypothesis-generation), and results (from experiment-design).
Structure
Use the standard IMRaD skeleton:
- Abstract — question, method, key result, conclusion. Written last, sharpest first sentence.
- Introduction — the gap: what is known, what is missing, and why it matters. End with the specific question/hypothesis.
- Methods — enough detail that the work is reproducible: design, measures, analysis. What was pre-registered vs exploratory.
- Results — the evidence, in order of the claims. Figures and tables carry the load; prose narrates. Report the failures and null results too.
- Discussion — interpretation: what this means, limits of the evidence, threats to validity, next steps.
- Conclusion — the takeaway, two to three sentences.
Writing Rules
- Every claim is backed by a result, a citation, or marked as interpretation.
- Statistical reporting includes effect size and uncertainty, never just p-values.
- Clearly separate findings from interpretation; do not bury caveats.
- Title the work as a question or result, not a metaphor.
- Credit prior work in citations; do not overstate novelty.
Output
research/papers/<slug>/paper.md — the manuscript.research/papers/<slug>/figures/ — figures referenced by the manuscript.- Side-by-side maintenance: if results change, update methods and results together, never in isolation.
Tone
Precise, modest, concrete. Favor the active voice and short sentences. Avoid hype; the science speaks.
