radiology-reporting
Select/audit reporting guidelines across imaging AI, diagnosis, reviews, guidelines and economics.
Install / Use
npx skills add huang-sir1/radiology-skills --skill radiology-reportingInstalls into whichever agent you are using.
SKILL.md
Installable skill definition
Quality Score
Category
LegalSupported Platforms
Our assessment of radiology-reporting
radiology-reporting scores 84/100 on our quality scale, 131st of 208 Legal skills we index.
Its SKILL.md is 14 KB long, well organised into 10 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.
Maintenance, license and trust
- The repository was last updated about 2 months ago, so radiology-reporting 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 foundOur 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-reporting compared with similar skills
All 4 of these similar skills score higher than radiology-reporting; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| radiology-reporting (this skill)by huang-sir1 | 84 | 1.8k | 56d ago | SKILL.md |
| x402by browser-use | 100 | 121.0k | 3d ago | SKILL.md |
| algorithmic-artby anthropics | 100 | 177.9k | 55d ago | SKILL.md |
| pptxby anthropics | 100 | 177.9k | 55d ago | SKILL.md |
| designby nextlevelbuilder | 100 | 149.0k | 6d ago | SKILL.md |
Frequently asked questions
- How do I install radiology-reporting?
- Run
npx skills add huang-sir1/radiology-skills --skill radiology-reporting. The install tabs above show the steps for each supported agent. - Which AI agents does radiology-reporting 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-reporting 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-reporting still maintained?
- The repository was last updated about 2 months ago, so radiology-reporting is actively maintained.
Skill content
View source on GitHubname: radiology-reporting description: "Select/audit reporting guidelines across imaging AI, diagnosis, reviews, guidelines and economics."
Radiology Reporting-Guideline Compliance
Use this skill to make an imaging study reviewer-proof on reporting. Radiology and the RSNA family require the relevant EQUATOR checklist at submission, and imaging-AI / radiomics papers are now judged against a specific, version-sensitive stack of guidelines. This skill (1) identifies the study type, (2) selects the correct guideline(s), (3) audits the manuscript item-by-item, and (4) returns a submission-ready checklist plus a prioritised fix list.
Core stance
- The checklist is the contract. A reviewer maps your paper to a guideline; do the same first, in their seat.
- Report honestly. Mark each item
PRESENT,PARTIAL, orMISSING. Never label something compliant to be agreeable. AMISSINGflag you surface is cheaper than a reviewer finding it. - Cite the location. Every
PRESENTclaim must point to a section / page / figure / supplement. If you cannot point to it, it isPARTIALat best. - Versions matter. Use the current version (CLAIM 2024 Update, TRIPOD**+AI** 2024,
CLEAR 2023, METRICS 2024, STARD-AI 2025 for AI accuracy studies, CONSORT/SPIRIT 2025,
PROBAST**+AI** 2025 — full list with sources in
references/guideline-versions.md). Name the version you audited against. - Provenance matters. Record the official guideline/checklist source, version/update date, access date, and any target-journal overlay. Treat remembered requirements as provisional.
- Reporting ≠ quality ≠ risk-of-bias. CLEAR (reporting) → METRICS / RQS (methodological quality) → PROBAST+AI / QUADAS-3 v1.2 (risk of bias). Different tools, different jobs; pick the right one(s).
- Don't invent the science. This skill audits reporting; it never fabricates the missing experiment, metric, or dataset. It tells the author what to add.
- Venue changes the stack, not the rigor. Radiology-family submissions stop at the
guideline checklist; Nature-portfolio submissions add a Reporting Summary / Editorial
Policy Checklist on top of the same guideline stack (→
nature-reporting-summary.md) — never treat the Reporting Summary as a replacement for CLAIM/TRIPOD+AI/CLEAR.
When to use
- "Which checklist does my study need?" / "What will Radiology require at submission?"
- "Audit this manuscript against CLAIM / TRIPOD+AI / STARD / CLEAR / METRICS / RQS."
- "Is my radiomics pipeline reported reproducibly (IBSI)?"
- "Fill in the CLAIM checklist with page numbers."
- "What's my risk-of-bias exposure under PROBAST+AI / QUADAS-3?"
- "Which checklist applies to a guideline/consensus, qualitative/mixed-methods, economic evaluation, implementation study or PPI report?"
- Pre-submission self-audit, or triaging a reviewer comment that cites a guideline.
Routing — pick the guideline(s) before auditing
Most imaging-AI papers need two or more of these (a reporting guideline and a quality/risk-of-bias tool).
| Study type | Primary reporting guideline | Add for quality / risk-of-bias |
|---|---|---|
| AI/ML system in medical imaging (any task) | CLAIM 2024 | TRIPOD+AI if it is a prediction model; DECIDE-AI for early clinical decision-support |
| Diagnostic/prognostic prediction model (incl. ML/DL) | TRIPOD+AI (2024) (+ TRIPOD-Cluster, TRIPOD+AI for Abstracts) | PROBAST / PROBAST+AI (2025) (risk of bias) |
| Radiomics (hand-crafted features → model) | CLEAR (2023) for reporting | METRICS (2024) and/or RQS / RQS 2.0 for quality; IBSI for feature reproducibility |
| Diagnostic accuracy (test vs reference standard) | STARD 2015; STARD-AI (2025) when the index test is AI | QUADAS-3 v1.2 when primary-study accuracy estimates are appraised in a review |
| DTA systematic review / meta-analysis | PRISMA-DTA (2018) | QUADAS-3 v1.2 per selected accuracy estimate; for within-study comparisons of two or more index tests, apply QUADAS-C alongside QUADAS-3 using the QUADAS-3 E&E adaptation guidance |
| Systematic review / meta-analysis (general) | PRISMA 2020 | AMSTAR-2; ROBIS |
| Scoping review / evidence map | PRISMA-ScR (2018) | critical appraisal only if performed and justified; AMSTAR-2/ROBIS do not replace the reporting checklist |
| Systematic-review or meta-analysis protocol | PRISMA-P (2015) | protocol reporting only; registration/freeze and review conduct remain with radiology-systematic-review |
| Living systematic review | PRISMA 2020 + PRISMA-LSR (2024) | update surveillance, changed-study ledger, transition/retirement and tailored flow |
| Review of outcome measurement instruments | PRISMA-COSMIN for OMIs (2024) | only when measurement properties/instruments are the review object; use design-matched COSMIN appraisal |
| Observational (cohort/case-control/cross-sectional) | STROBE | (REMARK for tumour-marker prognostic studies) |
| Randomised trial of an imaging/AI intervention | CONSORT 2025 (+ CONSORT-AI 2020 extension) | protocol: SPIRIT 2025 (+ SPIRIT-AI 2020) |
| Imaging biomarker / quantitative imaging | QIBA Profile reporting + STARD/TRIPOD as applicable | IBSI; phantom/repeatability (QIBA) |
| Practice guideline / biomedical consensus | RIGHT 2017 / ACCORD 2024 | AGREE II appraises guideline quality; recommendation methods stay with radiology-consensus-guideline |
| Qualitative / mixed methods | COREQ for interviews/focus groups or SRQR for broader qualitative work; add quantitative design guideline and integration reporting | method credibility/integration stay with radiology-qualitative-mixed-methods |
| Health economic evaluation | CHEERS 2022 (+ current CHEERS-AI when applicable) | economic estimand/model/uncertainty stay with radiology-health-economics |
| Implementation study | StaRI 2017 + design-specific guideline(s) | implementation determinants, strategies/outcomes and success stay with radiology-translation |
| Patient/public involvement report | GRIPP2 2017 | actual involvement/decision impact stay with radiology-design |
Open
references/guideline-router.mdfor the full decision tree, including hybrid studies (e.g. a radiomics prediction model validated for diagnostic accuracy → CLEAR + TRIPOD+AI + STARD + IBSI) and Nature-portfolio venues (add the Reporting Summary on top of whichever stack applies).
When to open extra files
| File | Open when | |---|---| | references/guideline-router.md | Choosing guideline(s); hybrid/edge-case study types; how guidelines stack; Nature-portfolio add-on; FUTURE-AI; TRIPOD-LLM | | references/claim-2024.md | Auditing a medical-imaging AI paper against the CLAIM 2024 Update | | references/tripod-ai-probast.md | Prediction-model reporting (TRIPOD+AI) and PROBAST / PROBAST+AI risk-of-bias | | references/clear-metrics-rqs.md | Radiomics reporting (CLEAR) and quality scoring (METRICS, RQS / RQS 2.0) | | references/ibsi-features.md | Making radiomic features reproducible/standardised (IBSI image processing + feature nomenclature) | | references/stard-prisma-quadas.md | Diagnostic-accuracy reporting (STARD), DTA reviews (PRISMA-DTA), and current risk-of-bias/applicability assessment (QUADAS-3 v1.2; QUADAS-C alongside it for comparative accuracy) | | references/prisma-general-scoping.md | General reviews (PRISMA 2020), protocols (PRISMA-P), scoping reviews/evidence maps (PRISMA-ScR), search reporting (PRISMA-S), living reviews (PRISMA-LSR), and measurement-instrument reviews (PRISMA-COSMIN for OMIs) | | references/radiology-submission-map.md | Mapping checklist items to where they belong in a Radiology manuscript + submission logistics | | references/nature-reporting-summary.md | Target is a Nature-portfolio journal — completing the Reporting Summary / Editorial Policy Checklist alongside the primary guideline stack | | references/guideline-consensus-qualitative-economic-implementation.md | Guideline/consensus, qualitative/mixed-methods, health-economics, implementation or PPI reporting; distinguishes reporting from method/appraisal/success |
Workflow
- Classify the study. Determine task (classification / detection / segmentation / prediction / diagnostic accuracy / discovery), data provenance, whether a model is developed and/or validated, whether the endpoint is accuracy, prognosis, biology, implementation or economics, and whether the output is a guideline/consensus or qualitative/mixed-method report.
- Select and verify guideline(s) from the routing table. State the official source, version/update date, and access date. If the study is hybrid, select the stack and say why each applies.
- Load the relevant reference file(s) and audit every item. For each item record:
Item ID | Requirement (short) | Status (PRESENT/PARTIAL/MISSING/NA) | Location | Fix | Owner | Closure evidence. - Prioritise fixes. Group into
Blocker(will trigger major revision / desk reject),Should-fix(reviewer will likely ask),Polish. Tie each blocker to the specific reviewer risk. - Cross-check integrity hot-spots (see below) — the items reviewers weaponise most.
- If the target is a Nature-portfolio venue, also complete the Reporting Summary /
Editorial Policy Checklist (
nature-reporting-summary.md) — additive, not a substitute. - Return the filled checklist + a one-screen executive summary + the prioritised fix
list. Offer to draft the missing text/Methods sentences (hand off to
radiology-writing).
Integrity hot-spots (audit these even if not asked)
These are the recurring reasons imaging-AI/radiomics papers get rejected:
- Data leakage / partition hygiene. Train/validation/test split made at the patient level (not slice/lesion); no test-set tuning; preprocessing, feature selection, harmonisation, and normalisation fit on training data only; augmentation never crosses the split. (CLAIM, TRIPOD+AI, METRICS, CLEAR all probe this.)
- External / independent validation. Internal CV alone is weak. State the validation type (internal resampling, temporal, geographic, fully external) and cohort source.
- Reference standard & ground truth. Who labelled, how many readers, expertise, blinding, adjudication, and the reference standard's own accuracy. (STARD, CLAIM.)
- Class/prevalence & spectrum. Report disease prevalence; flag artificial 1:1 sampling; describe the clinical spectrum (STARD spectrum bias; QUADAS-3 Participants domain).
- Radiomics reproducibility. Software + version, image preprocessing (resampling, discretisation/bin width, intensity normalisation), segmentation method and inter-observer reproducibility (ICC), feature definitions IBSI-compliant, and scanner/protocol harmonisation (e.g. ComBat). (CLEAR, METRICS, IBSI.)
- Sample size / EPV. Events-per-variable, or a stated sample-size rationale (Riley et al. for prediction models). High-dimensional features vs. n is the classic overfitting trap.
- Metrics match the task & prevalence. AUC alone is insufficient; report calibration and
clinical-utility (decision-curve) for prediction models; report CIs everywhere.
(Hand off computation to
radiology-stats.) - Code / model / data availability. Statement present and specific. (Hand off to
radiology-data.)
Output contract
Return, in this order:
- **`S
Truncated for display — read the full file on GitHub.
Related Skills
x402
121.0kSet up Browser Use Cloud payments with x402 — pay per request from a crypto wallet (USDC on Base mainnet), no signup or API key.
algorithmic-art
177.9kCreating algorithmic art using p5.js with seeded randomness and interactive parameter exploration. Use this when users request creating art using code, generative art, algorithmic art, flow fields, or particle systems.
pptx
177.9kUse this skill any time a .pptx or .potx file is involved in any way — as input, output, or both. This includes: creating slide decks, pitch decks, or presentations; reading, parsing, or extracting text from any .pptx or .potx file (even if the extracted content will be used elsewhere, like in an em…
design
149.0kComprehensive design skill: brand identity, design tokens, UI styling, logo generation (55 styles, Gemini, Atlas Cloud, or MuAPI AI), corporate identity program (50 deliverables, CIP mockups), HTML presentations (Chart.js), banner design (22 styles, social/ads/web/print), icon design (15 styles, SVG…
Languages
Trust signals
From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.
