SkillAgentSearch skills...

radiology-acquisition-qc

Audit imaging acquisition, reconstruction, series eligibility, quantitative transforms and protocol drift; not model training.

Install / Use

npx skills add huang-sir1/radiology-skills --skill radiology-acquisition-qc

Installs into whichever agent you are using.

About this skill
📄

SKILL.md

Installable skill definition

Quality Score

85/100

Category

Operations

Supported Platforms

Universal

Our assessment of radiology-acquisition-qc

radiology-acquisition-qc scores 85/100 on our quality scale, 521st of 746 Operations skills we index.

Its SKILL.md is 11 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.

Substance
29/30
Structure
13/20
Description
15/15
Adoption
14/20
Freshness
15/15

Maintenance, license and trust

  • The repository was last updated about 2 months ago, so radiology-acquisition-qc is actively maintained.
  • Our last check on 2026-11-04 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 2026-10-02. Automated pattern scan on 2026-10-02. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.

radiology-acquisition-qc compared with similar skills

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

SkillScoreStarsUpdatedFormat
radiology-acquisition-qc (this skill)by huang-sir1851.8k56d agoSKILL.md
taste-skillby Leonxlnx100116.8k8d agoCLAUDE.md
LocalAIby mudler10050.4ktodayMCP Server
x402by browser-use100121.0k3d agoSKILL.md
algorithmic-artby anthropics100177.9k55d agoSKILL.md

Frequently asked questions

How do I install radiology-acquisition-qc?
Run npx skills add huang-sir1/radiology-skills --skill radiology-acquisition-qc. The install tabs above show the steps for each supported agent.
Which AI agents does radiology-acquisition-qc 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-acquisition-qc 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-acquisition-qc still maintained?
The repository was last updated about 2 months ago, so radiology-acquisition-qc is actively maintained.

name: radiology-acquisition-qc description: "Audit imaging acquisition, reconstruction, series eligibility, quantitative transforms and protocol drift; not model training."

Imaging Acquisition, Reconstruction, and Measurement QC Router

Own the validity of the image as a measurement before annotation, feature extraction, model training, biological interpretation, or clinical deployment. Route to the smallest relevant modality playbook, freeze what was actually acquired and reconstructed, and return the strongest claim the verified measurement chain permits.

When to use

  • "Can these acquisition/reconstruction parameters support a quantitative study?" / "这批影像的采集重建参数能不能支撑定量研究?" Series/phase/sequence selection and exclusion.
  • "Can multi-center, multi-scanner, multi-protocol data be analysed together?" / "多中心、多扫描仪、多协议的数据能不能合并分析?"
  • "How do I interpret artifacts, dose, phantom or test-retest results?" / "伪影、剂量、重复扫描结果怎么解释?"
  • Freeze the measurement passport and claim ceiling before annotation, radiomics or model training.

Ownership

This skill owns:

  • acquisition, reconstruction and post-processing provenance;
  • intended series/phase/sequence/view/tracer selection and exclusion rules;
  • modality-specific artifacts, physical confounders and quantitative-unit validity;
  • phantom, test-retest and reconstruction-perturbation design;
  • protocol/site/scanner/software shift and the acquisition-side harmonisation boundary;
  • an Imaging measurement passport and PASS / CONDITIONAL / STOP measurement claim ceiling.

It does not own disease eligibility or treatment rules, de-identification, ROI annotation, radiomics/DL execution, statistical inference, clinical reader studies, mechanism interpretation, or individual-patient care.

Modes

| Mode | Use when | Output emphasis | |---|---|---| | plan | Designing acquisition/QC for a proposed study | minimum defensible protocol, stronger option, repeatability and shift plan | | audit | Images, DICOM exports, protocols, Methods or QC artifacts exist | provenance gaps, series-selection bias, artifacts, sensitivity and claim repair | | mentor | A learner has a clear acquisition/measurement question | decision ladder, physical rationale, falsifier, teach-back or transfer prompt | | interpret | Real QC, phantom, test-retest or protocol-shift results are supplied | magnitude, uncertainty, failure boundary and surviving measurement claim | | writing-handoff | Acquisition science is frozen | exact Methods/Results/Supplement fields, evidence state and prohibited upgrade |

Advisory and tutoring work is read-only. Create or update a passport only when the user requests a durable artifact.

Progressive routing

Resolve the active modality and measurement decision before opening references. Load only the matching file or files:

| Active object | Load | |---|---| | CT, dual-energy/spectral CT, photon-counting CT, CT perfusion or CT-derived quantitative measurement | references/ct-acquisition-reconstruction.md | | Structural/contrast-weighted MRI, DWI/ADC, DCE/DSC/ASL, quantitative mapping or MR reconstruction | references/mr-acquisition-reconstruction.md | | MRS or MRSI, including metabolite ratios, fitted peaks or concentration estimates | MR core plus references/mr-spectroscopy.md | | Task or resting-state BOLD fMRI, activation, connectivity or network measurement | MR core plus references/functional-mri.md | | DTI, DKI, NODDI, fixel analysis, tractometry or tractography | MR core plus references/diffusion-tensor-tractography.md | | PET, PET/CT, PET/MR, SPECT, SPECT/CT, SUV, kinetic or count-based measurement | references/pet-spect-acquisition-quantification.md | | Ultrasound, Doppler, CEUS, strain or shear-wave elastography | references/ultrasound-ceus-elastography.md | | Projection radiography, mobile radiography, mammography, DBT or synthetic 2D mammography | references/projection-radiography-mammo-dbt.md | | Phantom, test-retest, longitudinal change, multicentre protocol shift, harmonisation or reconstruction sensitivity | references/phantom-test-retest-protocol-shift.md | | Fluoroscopy, DSA/angiography, cone-beam CT, image-guided or interventional imaging | LOCAL_EXTENSION_REQUIRED; do not substitute the CT or projection-radiography playbook |

For PET/CT load PET/SPECT plus CT; for PET/MR load PET/SPECT plus MR. For multimodal studies, load only each modality that contributes to the requested measurement claim. Add the common phantom/shift playbook when repeatability, longitudinal, cross-scanner, cross-site or absolute quantitative equivalence is material.

Fluoroscopy, DSA/angiography, cone-beam CT and interventional imaging have time-dependent dose, device, subtraction, geometry, procedural and safety contracts not represented here. Mark the route LOCAL_EXTENSION_REQUIRED and stop acquisition-specific assurance until a dedicated, locally verified playbook and responsible radiologist/physicist/procedural owner are supplied. Ordinary CT or projection-radiography rules cannot be borrowed as a substitute.

Required routing workflow

  1. Name the requested measurement object and claim: visual adequacy, relative signal, absolute quantity, longitudinal change, cross-site comparability, reconstruction gain, or downstream input fitness.
  2. Identify the independent hierarchy: patient -> examination/time point -> series/acquisition -> reconstruction/derived object -> frame/slice/voxel/measurement. Do not treat reconstructions, views, frames or repeated scans as independent patients.
  3. Load the modality playbook and reconstruct the acquisition-to-measurement chain from supplied artifacts. Mark unsupported fields AUTHOR_INPUT_NEEDED; do not infer settings from a series description or filename.
  4. Freeze the series/phase/sequence/view/tracer and reconstruction inclusion rule before outcome, model-performance or mechanism results are used to select inputs.
  5. Add the common playbook when a repeatability or transport claim is intended. Distinguish a represented protocol/scanner level from a wholly unseen level and untouched evaluation from adaptation.
  6. Assign PASS, CONDITIONAL or STOP to the exact measurement claim, then hand off the passport without upgrading its evidence state.

Use templates/imaging-measurement-passport.md for a requested persistent/full audit artifact.

Claim ceiling

  • PASS: the acquisition, reconstruction, series rule, QC, units and applicable repeatability/ transport evidence are traceable and fit the stated claim.
  • CONDITIONAL: the measurement remains usable for a narrower population, protocol, relative comparison or sensitivity-bounded claim, with the unresolved source named.
  • STOP: the central measurement is not identifiable or reproducible because the wrong/unknown series, invalid units, result-dependent input selection, dominant unbroken acquisition confounding, failed QC, or an unsupported quantitative transform determines the claim.

A standardized protocol is not proof of cross-platform equivalence. Phantom repeatability is not clinical validity. Harmonisation is not a substitute for missing acquisition metadata, invalid quantification or an outcome-confounded site. A derived map or model output is not a direct physical measurement merely because it is stored as an image.

Handoffs

| Requested decision | Unique owner | Send from this skill | |---|---|---| | disease pathway, endpoint, reference standard, treatment timeline or current clinical rule | radiology-clinical-domain | measurement passport, protocol limits, artifacts and claim ceiling | | cohort inventory, DICOM de-identification, release or repository plan | radiology-data; governance to radiology-ethics | series/object inventory and sensitive-data flags without asserting authorization | | ROI/VOI or reader annotation procedure | radiology-annotation | eligible series, geometry, resolution and artifact constraints | | radiomics preprocessing/extraction | radiology-radiomics | accepted input series, units, reconstruction/shift and repeatability fields | | imaging model/reconstruction network training | radiology-deep-learning | measurement target, input provenance, paired reference and protected evaluation rule | | parameter comparison, ablation or method-fit judgement | radiology-method-evaluation | frozen parameter ledger and acquisition claim; numerical inference to radiology-stats | | radiologist reader study, threshold-to-action or prospective workflow | radiology-translation | intended image presentation, failure/nondiagnostic states and measurement ceiling | | imaging-to-pathology/omics mechanism | radiology-radiogenomics | physical contrast candidates, measurement state and technical alternatives |

Output contract

Return only the parts needed for the selected mode:

  1. Route and evidence state — mode, modality/subtype, requested measurement and supplied artifacts.
  2. Measurement chain — acquisition -> reconstruction -> post-processing -> selected object -> value.
  3. Series and hierarchy lock — included/excluded objects, unit and repeated-measure structure.
  4. Coverage and disposition denominators — eligible, acquired, exported/received, QC-assessed, accepted, nondiagnostic and excluded counts at the declared grain, with reconciled reasons.
  5. QC and physical-confounder map — artifact, likely effect, evidence, test and disposition.
  6. Repeatability/shift read — when applicable, represented/unseen levels, raw primary analysis, adaptation status and sensitivity.
  7. Claim verdict — PASS / CONDITIONAL / STOP, allowed wording, prohibited upgrade and smallest next action.
  8. Handoff — exact fields and unresolved assumptions for the unique receiving owner.

Boundaries

  • Do not infer a scanner setting, dose, contrast/tracer exposure, sequence, reconstruction, QC pass or quantitative unit from convention.
  • Do not prescribe or alter an individual patient's examination, contrast, radiation, diagnosis or treatment. Local protocol approval, radiation safety and equipment acceptance remain with qualified radiologists, technologists, medical physicists and institutional governance.
  • Do not run a reconstruction, modify source images, or persist a passport unless the requested outcome authorizes that action.
  • Exact standards, profiles and practice parameters are versioned and jurisdiction-dependent. Verify the current authoritative artifact and applicability before asserting conformance.
  • Fluoroscopy, DSA/angiography, cone-beam CT and interventional imaging are LOCAL_EXTENSION_REQUIRED; this skill has no dedicated contract for them and must not upgrade a generic CT or X-ray review into modality- or procedure-specific assurance.

Related Skills

View on GitHub
GitHub Stars1.8k
CategoryOperations
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