SkillAgentSearch skills...

radiology-writing

Draft/restructure manuscripts or theses from fixed evidence; not polishing, review or rebuttal. CN: 论文写作、英文稿、结构化摘要

Install / Use

npx skills add huang-sir1/radiology-skills --skill radiology-writing

Installs into whichever agent you are using.

About this skill
📄

SKILL.md

Installable skill definition

Quality Score

84/100

Supported Platforms

Universal

Our assessment of radiology-writing

radiology-writing scores 84/100 on our quality scale, 860th of 1,153 Content & Media skills we index.

Its SKILL.md is 19 KB long, well organised into 9 sections and no code examples: a thorough specification that gives an agent plenty to work with.

With 1,785 GitHub stars, it is one of the more widely adopted skills in the catalogue.

Substance
30/30
Structure
13/20
Description
12/15
Adoption
14/20
Freshness
15/15

Maintenance, license and trust

  • The repository was last updated about 2 months ago, so radiology-writing is actively maintained.
  • Our last check on 2026-11-05 found the source still online.
  • No license is declared. By default that means all rights are reserved: you can read it, but reusing or redistributing it is not clearly permitted. Ask the author before building on it commercially.
  • Its trust signals score 88/100, with 1 caution from licensing, adoption, age or documentation. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.

Safety scan

No issues found

Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful.

AI review by kimi-k2.7-code on 2026-10-01. Automated pattern scan on 2026-10-01. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.

radiology-writing compared with similar skills

All 4 of these similar skills score higher than radiology-writing; compare them before choosing.

SkillScoreStarsUpdatedFormat
radiology-writing (this skill)by huang-sir1841.8k56d agoSKILL.md
siyuanby siyuan-note10047.3ktodayMCP Server
x402by browser-use100121.0k3d agoSKILL.md
algorithmic-artby anthropics100177.9k55d agoSKILL.md
pptxby anthropics100177.9k55d agoSKILL.md

Frequently asked questions

How do I install radiology-writing?
Run npx skills add huang-sir1/radiology-skills --skill radiology-writing. The install tabs above show the steps for each supported agent.
Which AI agents does radiology-writing work with?
It is written for Universal, as a SKILL.md file. Other agents that read the same format can often use it too.
Is radiology-writing safe to use?
Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful. It declares no license and scores 88/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
Is radiology-writing still maintained?
The repository was last updated about 2 months ago, so radiology-writing is actively maintained.

name: radiology-writing description: "Draft/restructure manuscripts or theses from fixed evidence; not polishing, review or rebuttal. CN: 论文写作、英文稿、结构化摘要"

Venue-routed imaging and mechanism manuscript writing

Use this skill to construct imaging or mechanism-research prose — argument first, then sentences — in the selected venue's real shape. Radiology is the fallback only for an imaging-centered manuscript whose venue is not yet selected; it is not the default shape for a standalone bulk RNA, single-cell, spatial or perturbation paper. For sentence-level polish use radiology-polishing.

Core stance

  • Author evidence first. Never invent results, metrics, p-values, cohort numbers, citations, mechanisms, or limitations. Missing input → explicit placeholder or a question.
  • Write the argument before the sentences. One-sentence claim → section architecture → paragraph jobs → prose.
  • Match the selected venue family. Use Radiology Summary/Key Results only on the RSNA route; Nature/npj, Cell Press, JAMA Network, Lancet and Wiley routes are not interchangeable.
  • Preserve the scientific packet. When a validated radiogenomics handoff exists, stable Claim IDs, modality roles, independent unit, matched n, evidence state, claim branch/verdict/ceiling and protected placement are immutable. Return conflicts to the scientific ledger.
  • Preserve the evidence-synthesis packet. For evidence-synthesis, consume the frozen protocol, search, selection-flow, extraction, risk-of-bias/applicability, synthesis and certainty artifacts from radiology-systematic-review. Preserve the study-family/report/effect-row hierarchy and the synthesized claim ceiling; this route does not require a radiogenomics handoff.
  • Bound the claim. Ambitious but evidence-bounded; calibrate verbs to evidence (demonstrate → suggest → may reflect).
  • Reporting-aware. Every Methods/Results element should satisfy the relevant checklist item (CLAIM/TRIPOD+AI/CLEAR/STARD) — cross-check with radiology-reporting.

When to use

  • Draft/rebuild any section: title, abstract, Introduction, Methods, Results or Discussion, plus venue-specific front matter such as Summary statement/Key Results only when the selected route requires it.
  • Turn Chinese lab notes / mixed drafts into submission-ready English.
  • Chinese→English translation/reconstruction of manuscript text (language translation — not to be confused with radiology-translation, which is clinical/bench-to-bedside translation). Translate the argument and intent, not clause order; open references/chinese-author-workflow.md first.
  • Restructure a rejected draft to the Radiology argument shape.
  • Build or audit a monograph thesis, article-based/cumulative thesis, chapter package, thesis-to-paper conversion or whole-thesis argument spine from fixed evidence. Institutional regulations and examiner decisions remain external human authority.
  • 中文高频说法:“帮我把中文实验记录/组会思路写成英文投稿稿。”“按 Radiology 风格重构摘要和讨论。”

When to open extra files

| File | Open when | |---|---| | references/article-architecture.md | Section order, argument flow, and the Radiology manuscript skeleton | | references/structured-abstract.md | Writing the structured abstract + Summary statement + Key Results box | | references/methods.md | Materials and Methods for imaging/AI/radiomics studies (what must appear, in order) | | references/results.md | Results narrative: flow, performance with CIs, comparisons, validation | | references/discussion.md | Structured Discussion (key-finding first → context → limitations → conclusion) | | references/chinese-author-workflow.md | Notes are Chinese / mixed / lab-note style; translate intent and argument, not clause order | | references/nature-family-shape.md | Target is Nature Medicine / Nature Biomedical Engineering / Nature Communications / npj Digital Medicine, etc. — unstructured abstract, no Summary statement/Key Results box, different Methods placement | | references/cell-jama-wiley-shape.md | Target is Cancer Cell, Cell Reports Medicine, JAMA Network Open or Advanced Science; keeps the families distinct and fails closed when the exact shape is unverified | | references/argument-spine-and-stage-gates.md | Full manuscript rebuild, rejected-paper rescue, unclear contribution, high-impact submission, or a draft whose story/figures/results do not yet lock together | | references/journal-family-writing-style.md | Target journal family is known, or the user supplied author-guide PDFs/classic papers and wants the manuscript to carry that venue's writing style | | references/section-contract-and-style-profile.md | Full-section drafting, multiple coauthors, author-voice calibration, or when prose keeps drifting beyond the allowed evidence | | ../radiology-radiogenomics/references/scientific-handoff-contract.md | A prior radiogenomics/mechanism task supplies a shared-state packet; validate it and return a Claim-ID drift ledger | | ../radiology-method-evaluation/references/parameter-and-methodology-evaluation.md | Drafting or placing parameter provenance, metric/method comparison, ablation, sensitivity, robustness, failure or efficiency evidence; do not convert a planned/unverified evaluation into Results prose | | ../radiology-systematic-review/references/review-methods-and-quality-gates.md | Drafting a systematic/scoping review or meta-analysis from a fixed writing handoff; verify the protocol/search/flow/extraction/RoB/synthesis/certainty chain and keep study families distinct from reports/effect rows | | references/thesis-and-dissertation-architecture.md | Monograph or article-based thesis, chapter drafting/revision, thesis-to-paper conversion, synthesis chapter, whole-thesis audit or defense handoff |

Intake (identify before drafting)

  • Target venue/shape: resolve RSNA, Nature/npj, Cell Press, JAMA Network, Advanced/Wiley, European, NEJM, Science or Lancet before finalising front matter. For Cancer Cell, Cell Reports Medicine, JAMA Network Open or Advanced Science, open cell-jama-wiley-shape.md. If venue is undecided, use a reversible journal-neutral scientific spine; use the Radiology fallback only for an imaging-centered paper.
  • Scientific state: if a mechanism or radiogenomics packet is supplied, record its path, canonical packet_sha256, distinct physical file SHA-256 when transported, source-manifest digest and writing_handoff_status; validate it before consuming W-HANDOFF-READY. If no packet exists, build a local map of scope, modality roles, independent unit, usable n, material Claim IDs, evidence locations, evidence states and wording ceilings from the supplied artifacts. Mark unresolved fields AUTHOR_INPUT_NEEDED; do not block traceable sections merely because another module was not invoked first. An untraceable Results or mechanism claim receives a visible scaffold or W-RETURN-TO-LEDGER, never invented science.
  • Evidence-synthesis state: for study_scope=evidence-synthesis, record the review route, lifecycle, protocol/search/selection-flow/extraction/risk-of-bias-applicability/synthesis/certainty artifact IDs, versions and SHA-256s, study-family hierarchy, pooling decision, certainty and supported/forbidden wording. If a decision-bearing artifact is missing or not fixed, write only a visible Methods/Results scaffold and return the affected item to radiology-systematic-review.
  • Section(s) requested.
  • Study type: diagnostic-accuracy, prediction model, radiomics, radiogenomics, reader study, observational, trial, scoping review, systematic narrative synthesis, or meta-analysis.
  • Core claim: what the study actually shows.
  • Evidence: cohorts (n, source, dates), metrics with CIs, comparisons, validation.
  • Method-evaluation state: Evaluation IDs, parameter/configuration IDs, benchmark/ablation artifacts, selection/freeze points, evidence state and claim consequence when the paper makes a methodological or robustness claim.
  • Boundary: where the claim stops (single-centre? retrospective? prevalence?).
  • Limits: target word/figure counts (verify against current author instructions).
  • Thesis state when applicable: degree/programme, monograph versus article-based form, institutional rule source/date, included-paper status and reuse rights, chapter/version map, cross-chapter claim IDs, contribution statement, examination/defense milestone and embargo route.

If core claim, evidence, or boundary is missing, surface the gap and offer a scaffold with placeholders rather than inventing content.

Writing workflow

  1. For full manuscripts or high-impact rebuilds, open argument-spine-and-stage-gates.md and establish the project context, contribution-first gate, and results-as-validation map before drafting long prose.
  2. For radiogenomics or mechanism work, validate an existing packet and establish a Claim-ID drift ledger. Without a packet, create the bounded local scientific-state map described above and continue wherever evidence is traceable; escalate only affected scientific uncertainties to radiology-radiogenomics.
  3. For evidence synthesis, select study_scope=evidence-synthesis, consume the fixed radiology-systematic-review: writing-handoff, and verify all seven review-artifact roles plus the study-family/report/effect-row map before drafting Results or certainty language. Planned, author-reported or unresolved pooling output stays out of Results. Route a changed eligibility, extraction, bias, synthesis or certainty decision back to the review owner rather than creating a radiogenomics packet.
  4. For target-venue writing taste, open journal-family-writing-style.md after confirming the venue family; use it to adjust article shape, title/abstract rhythm, key points, and clinical relevance language.
  5. For parameter or methodology evidence, consume a completed radiology-method-evaluation matrix when available. If the evaluation is missing, separate Methods reporting facts from the unproven Results claim and return only the affected question to that module.
  6. For thesis/dissertation work, open thesis-and-dissertation-architecture.md; freeze the institution/programme rule passport and choose monograph, article-based, chapter-draft, thesis-to-paper or whole-thesis-audit. Keep a thesis-wide contribution spine, stable Claim IDs, included-paper/version/reuse-rights map and cross-chapter non-duplication ledger. Do not convert a defense outcome, examiner judgment or repository/embargo decision into a writing-model verdict.
  7. One-sentence argument: "In [population/modality], we show [advance] using [approach], supported by [key result with CI], with [boundary]."
  8. Pick the architecture conditionally — RSNA/Radiology route → article-architecture.md and, when applicable, structured-abstract.md; Nature/npj route → nature-family-shape.md; Cancer Cell/Cell Reports Medicine/JAMA Network/Advanced-Wiley route → cell-jama-wiley-shape.md. For other verified families, use the applicable current guide/style reference. If the family is unknown or its exact s

Truncated for display — read the full file on GitHub.

Related Skills

View on GitHub
GitHub Stars1.8k
CategoryContent
Updated1mo ago
Forks17

Languages

Python

Trust signals

88/100

From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.

1 medium