From lancet-skills
Structures and trims a Lancet Article main text (IMRaD, ~3000–3500 words) with a cautious Discussion, Research in context panel, and reference budget (~30).
How this skill is triggered — by the user, by Claude, or both
Slash command
/lancet-skills:lancet-writingThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
- Drafting or trimming a Lancet original-research **Article**.
The Lancet original-research Article uses standard IMRaD with the Research in context panel:
| Section | Holds |
|---|---|
| Introduction | The problem, the gap, and the specific objective — brief (the systematic search lives in the panel, not a long literature review). |
| Methods | Design, setting/countries, participants, randomisation/allocation/blinding (or observational design), outcomes, sample size, statistical analysis (pre-specified), registration, ethics. Methods are detailed and stay in the main text (clinical journals expect full Methods, unlike Science's supplement model). |
| Results | Recruitment/flow (with the flow diagram), Table 1, primary outcome, secondary outcomes, subgroups, harms — in a logical order. |
| Discussion | Principal findings → comparison with other studies → strengths and limitations → implications. |
| Research in context panel | The mandatory three-part box (see lancet-research-in-context). |
lancet-figures-tables); extended methods/results → appendix.Write the Discussion in this order, and keep it cautious:
The Lancet prizes caution, not overstatement. Match claims to design: an observational study shows association, not causation; a single trial rarely "proves" — it "supports" or "provides evidence for." A non-inferiority trial cannot claim superiority.
The Lancet's editors read for a globally minded, cautious voice: a brief Introduction (the systematic search lives in the panel), full Methods in the main text, and a Discussion judged on restraint — observational studies show association, a single trial supports rather than proves, a non-inferiority trial never claims superiority. Patterns reviewers flag: a literature-review Introduction; Methods pushed offstage; "proves"/superiority overstatement; an over-budget reference list. Confirm current limits in the author guidelines.
A hypothetical Lancet Article on a multi-country observational cohort.
Length audit (illustrative):
Main text 3 980 words -> over ~3 500 by ~480; trim Introduction and Discussion
References 47 -> over ~30; keep the load-bearing 30
Claim calibration:
DRAFT: "This proves the exposure causes the outcome."
FIX: "The exposure was associated with the outcome (adjusted HR 1.34,
95% CI 1.18-1.52); residual confounding cannot be excluded."
The fix trims to budget and recasts the overstatement as a calibrated association (95% CI).
【Sections present】 Introduction / Methods / Results / Discussion / Research-in-context panel — all? yes/no
【Methods in main text?】 yes (good) / pushed to appendix (FIX)
【Main-text word count】 N → over/under ~3000–3500 by M
【Reference count】 N → vs ~30 budget
【Discussion four moves】 principal findings / comparison / strengths-limitations / implications — all? yes/no
【Overstatement check】 causal language calibrated to design? yes/no
【Next】 lancet-ethics
npx claudepluginhub brycewang-stanford/awesome-journal-skills --plugin lancet-skillsStructures and tightens NEJM Original Articles into terse IMRAD format with short main text, limited references, claim-first results, and sober discussion with limitations.
Routes manuscript workflow for The Lancet: determines which lancet-* sub-skill to invoke based on current stage (fit, design, analysis, writing, rebuttal).
Structures a Cell Article with the required sections, length budget, and Results subheadings that drive the narrative arc.