[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-aws-labs-imaging-study-design":3,"mdc-qsjwd5-key":49,"related-repo-aws-labs-imaging-study-design":1251,"related-org-aws-labs-imaging-study-design":1348},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":12,"stars":26,"repoUrl":27,"updatedAt":28,"license":29,"forks":30,"topics":31,"repo":44,"sourceUrl":47,"mdContent":48},"imaging-study-design","design medical imaging studies and biomarkers","Reasoning skill for medical imaging study design and biomarker selection. Use when the user asks to plan an imaging study, choose a preprocessing strategy, select an imaging biomarker, design a radiomics pipeline, handle DICOM de-identification, plan longitudinal imaging analysis, or pick a registration target. Triggers include \"imaging study\", \"preprocessing strategy\", \"DICOM de-identification\", \"imaging biomarker\", \"radiomics\", \"longitudinal imaging\", \"registration target\", \"MNI vs native space\", \"scanner harmonization\", \"ComBat\", \"IBSI\", \"multi-site imaging\", \"burned-in PHI\", \"test-retest reliability\", \"volumetric biomarker\", \"diffusion MRI\", \"perfusion imaging\", \"fMRI study design\", \"spectroscopy biomarker\".",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},"aws-labs","AWS Labs","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Faws-labs.png","awslabs",[13,17,20,23],{"name":14,"slug":15,"type":16},"Research","research","tag",{"name":18,"slug":19,"type":16},"Healthcare","healthcare",{"name":21,"slug":22,"type":16},"Clinical Trials","clinical-trials",{"name":24,"slug":25,"type":16},"Life Sciences","life-sciences",4,"https:\u002F\u002Fgithub.com\u002Fawslabs\u002Fhcls-agent-skills","2026-07-12T08:38:01.637878",null,0,[32,33,34,35,36,37,38,39,40,25,41,42,43],"agent-skills","agentcore","ai-agents","amazon-quick-desktop","claims-processing","drug-discovery","genomics","healthcare-ai","kiro","medical-imaging","risk-adjustment","strands-agents",{"repoUrl":27,"stars":26,"forks":30,"topics":45,"description":46},[32,33,34,35,36,37,38,39,40,25,41,42,43],"Agent skills for healthcare and life sciences: genomics, imaging, claims, drug discovery, and more. Works with Amazon Quick, Kiro, Amazon AgentCore, AWS Strands SDK, Claude Code, Codex, and any Agent Skills-compatible platform.","https:\u002F\u002Fgithub.com\u002Fawslabs\u002Fhcls-agent-skills\u002Ftree\u002FHEAD\u002Fskills\u002Fimaging-study-design","---\nname: imaging-study-design\ndescription: Reasoning skill for medical imaging study design and biomarker selection. Use when the user asks to plan an imaging study, choose a preprocessing strategy, select an imaging biomarker, design a radiomics pipeline, handle DICOM de-identification, plan longitudinal imaging analysis, or pick a registration target. Triggers include \"imaging study\", \"preprocessing strategy\", \"DICOM de-identification\", \"imaging biomarker\", \"radiomics\", \"longitudinal imaging\", \"registration target\", \"MNI vs native space\", \"scanner harmonization\", \"ComBat\", \"IBSI\", \"multi-site imaging\", \"burned-in PHI\", \"test-retest reliability\", \"volumetric biomarker\", \"diffusion MRI\", \"perfusion imaging\", \"fMRI study design\", \"spectroscopy biomarker\".\nusage: Invoke when planning an imaging study, choosing preprocessing strategies, selecting imaging biomarkers, or designing radiomics pipelines.\nversion: 1.0.0\ntags: [skill, category:reasoning, medical-imaging, study-design, biomarker, radiomics, hcls]\n---\n\n# Imaging Study Design and Biomarker Selection\n\n## Overview\n\nThis skill encodes methodology for designing medical imaging studies and selecting imaging biomarkers. It guides decisions about preprocessing, registration targets, de-identification, biomarker choice, radiomics stability, and multi-site harmonization. It is a reasoning skill — it does not generate pipeline code. For concrete tooling (FreeSurfer, FSL, ANTs, PyRadiomics, nnU-Net, MONAI), pair this skill with a pipeline skill after the study design is settled.\n\nThe central principle: **preprocessing and biomarker choice must preserve the signal that answers the clinical question**. A pipeline that is standard for one question (e.g., MNI registration for group comparison) can destroy the signal for another (e.g., individual atrophy in longitudinal studies).\n\n## Usage\n\nInvoke this skill when planning an imaging study or critiquing an existing imaging pipeline. Typical prompts:\n\n- \"I want to measure hippocampal atrophy over 2 years in Alzheimer's patients — how should I preprocess?\"\n- \"Design a multi-site radiomics study for NSCLC prognosis.\"\n- \"We're collecting brain MRI across three scanners. What harmonization do I need?\"\n- \"Is it safe to share these DICOMs after standard de-identification?\"\n- \"Which imaging biomarker should I use for presurgical mapping in epilepsy?\"\n\nUse the Decision Framework at the end of Core Concepts to structure any imaging study question.\n\n## Response Format\n\n- Lead with the direct recommendation or classification (≤3 sentences)\n- Structure as: recommendation → justification (citing specific criteria\u002Fthresholds) → caveats\n- Use tables for comparisons; bullet points for criteria lists\n- Omit background the user already knows — they asked the question\n- Target: 200-400 words unless the user requests exhaustive detail\nThe decision trees and frameworks in this skill are for internal reasoning only. Apply them to reach your conclusion, but do not reproduce them in your response. Present only the final recommendation with supporting evidence.\n\n\n## Core Concepts\n\n### 1. Preprocessing strategy is downstream-goal-dependent\n\nThere is no universal \"preprocess MRI\" recipe. The pipeline must match the analysis target.\n\n| Downstream goal | Preprocessing priorities | What to avoid |\n| --- | --- | --- |\n| Group comparison (cross-sectional) | Register to common template (MNI\u002FSPM), smooth, modulate if VBM | Native-space-only analysis |\n| Longitudinal atrophy \u002F change | Rigid register follow-ups to subject baseline, then (optionally) baseline to template | Registering each timepoint independently to MNI |\n| Lesion segmentation | Bias field correction, skull strip, intensity normalization; registration often optional | Aggressive smoothing that blurs lesion edges |\n| Radiomics \u002F texture | Standardize acquisition first (voxel size, reconstruction, binning); resample to isotropic; discretize intensity deterministically | Mixing protocols without harmonization; default-bin radiomics on heterogeneous data |\n| Functional MRI | Slice-timing (if long TR), motion correction, distortion correction, coregister to anatomical, normalize if group, bandpass filter as appropriate | Normalizing before motion scrubbing; smoothing before ICA denoising |\n| Diffusion MRI | Denoise, Gibbs unringing, eddy + motion correction, bias correction, then tensor\u002FODF fitting | Reordering these steps; skipping distortion correction when reverse-PE is available |\n\n### 2. Registration target decisions\n\nThe registration target encodes an assumption about where the signal lives.\n\n- **MNI \u002F population template** — for cross-sectional group statistics, voxelwise analyses (VBM, TBSS), atlas-based ROI extraction. Accepts that individual anatomy is partly warped away.\n- **Subject baseline (within-subject template)** — for longitudinal analyses of change (atrophy, tumor growth, lesion evolution). Preserves each subject's anatomy and measures change in a consistent individual frame.\n- **Native space** — for lesion analysis, surgical planning, radiomics on tumors, any case where registration would deform the structure of interest or introduce interpolation artifacts in the ROI.\n- **Subject anatomical (T1)** — for within-subject cross-modal alignment (fMRI, DWI, PET to T1) before any group normalization.\n\nHeuristic: ask \"if I warp this image, do I still measure the thing I care about?\" If the answer is no (volume change, lesion shape, tumor texture), keep the measurement in native or within-subject space.\n\n### 3. DICOM de-identification risk surface\n\nStandard tag-based de-identification (DICOM PS 3.15 Basic Application Level Confidentiality Profile) is necessary but not sufficient. Residual PHI risks:\n\n- **Burned-in annotations in pixel data** — ultrasound frames, screen captures, secondary captures, dose reports, 3D reformats. Patient name\u002FMRN\u002FDOB is often rendered into the pixels. Tag-level scrubbing does not remove this. Requires visual QC or OCR-based pixel redaction.\n- **Private tags by manufacturer** — Siemens CSA headers (`0029,xx10` \u002F `0029,xx20`) may contain protocol and sometimes patient info; GE private blocks (e.g., `0043,xx`) and Philips MR private tags can hold acquisition metadata that includes identifiers or free-text. Default DICOM profiles may retain or only partially scrub these. Decide per-vendor: strip all private tags unless a specific block is needed, and validate.\n- **Structured reports (SR) and encapsulated PDFs** — narrative findings often contain patient name, referring physician, dates, prior history. Must be treated as free-text and redacted or dropped.\n- **UIDs encoding dates or MRNs** — some sites generate Study\u002FSeries\u002FSOP Instance UIDs from timestamps or record numbers. Rewrite UIDs with a consistent hash (preserving referential integrity within the release) rather than leaving originals.\n- **Facial reconstructability from head MRI\u002FCT** — high-resolution T1\u002FCT allow face re-rendering. Apply defacing (e.g., remove or deform voxels around the face) for public release; consider skull-stripping for research-only shares.\n- **Acquisition dates and times** — even after name removal, exact dates can re-identify when combined with auxiliary data. Date-shift per subject with a preserved offset if relative timing matters.\n\nA defensible de-identification plan specifies: which profile is applied, how private tags are handled, whether pixel data is inspected, how SR\u002FPDF objects are handled, UID remapping strategy, defacing policy, and a QC sample checked by a human.\n\n### 4. Imaging biomarker selection by indication\n\nMatch the biomarker to the biology being probed.\n\n| Indication | Modality \u002F contrast | Biomarker | Rationale |\n| --- | --- | --- | --- |\n| Alzheimer's disease, MCI | Structural T1 MRI | Hippocampal volume, cortical thickness, ventricular volume | Measures neurodegeneration directly |\n| Multiple sclerosis | T2\u002FFLAIR + T1 | Lesion load, lesion count, brain parenchymal fraction, cord area | Disease burden and atrophy |\n| White matter disease, DAI, schizophrenia | Diffusion MRI | FA, MD, RD, AD (tract-based or voxelwise) | Microstructural integrity |\n| Acute stroke | DWI + PWI | DWI lesion volume, PWI-DWI mismatch, CBF | Core vs penumbra triage |\n| Brain tumor (glioma) | DSC\u002FDCE perfusion, MRS, diffusion | rCBV, Ktrans, ADC, NAA\u002FCho ratio | Grade, treatment response, pseudoprogression vs recurrence |\n| Presurgical planning (epilepsy, tumor) | Task and resting-state fMRI, DTI tractography | Language\u002Fmotor activation maps, tract-to-lesion distance | Functional localization |\n| Neuronal integrity (various) | 1H-MRS | NAA\u002FCr, Cho\u002FCr, mI\u002FCr | Metabolic markers of neuronal loss, membrane turnover |\n| Oncology outside brain | CT, PET\u002FCT, mpMRI | Volumetry, SUVmax\u002FSUVpeak, radiomic signatures, ADC | Staging, response (RECIST, PERCIST) |\n| Cardiac | Cine, LGE, T1\u002FT2 mapping | EF, strain, LGE burden, native T1, ECV | Function, fibrosis, edema |\n\nSelection criteria: (1) biological plausibility — is this biomarker mechanistically linked to the question? (2) measurability — does the available scanner\u002Fprotocol yield adequate SNR and reliability? (3) validity — has it been validated for this indication and this population? (4) practicality — can it be obtained within the study's scan time and cost?\n\n### 5. Radiomics stability and sample-size discipline\n\nRadiomics features are notoriously unstable. A signature built on unstable features will not generalize.\n\n- **Test-retest reliability** — require ICC ≥ 0.75 (commonly ICC ≥ 0.9 for high confidence) on repeat scans for any feature entering the model. Use a test-retest cohort or a publicly available one (e.g., RIDER Lung CT) aligned to the modality.\n- **IBSI compliance** — use an IBSI-compliant extractor (PyRadiomics with IBSI settings, CERR, LIFEx) and report IBSI feature definitions and hash. Non-compliant extractors produce features that cannot be reproduced across sites.\n- **Segmentation robustness** — evaluate inter-rater and intra-rater ICC for feature values under realistic contour variability, not just ideal contours.\n- **Acquisition standardization first** — fixing voxel size, reconstruction kernel, and intensity discretization (fixed bin width vs fixed bin count — choose one and justify) eliminates more variance than any post-hoc harmonization.\n- **Harmonization** — apply ComBat (or its variants: parametric\u002Fnon-parametric, with or without covariate preservation) to remove scanner\u002Fsite effects after acquisition standardization, not as a substitute for it.\n- **Feature reduction before modeling** — cluster redundant features (|ρ| > 0.8), drop unstable features, then apply a supervised reducer (LASSO, mRMR) inside cross-validation. Cap the final model at roughly **1 feature per 10 events** (EPV ≥ 10) to avoid overfitting. For low-event studies, report this constraint explicitly; it often means choosing 3–5 features, not 20.\n- **Report a TRIPOD- or CLEAR-style pipeline description** so the study is reproducible.\n\n### 6. Multi-site study considerations\n\nMulti-site studies trade statistical power for between-site heterogeneity. Plan for the heterogeneity.\n\n- **Protocol standardization** — adopt ADNI-style or equivalent harmonized protocols (sequence parameters, resolution, coverage, reconstruction). Agree on inclusion criteria for scanner models and software versions.\n- **Scanner and coil effects** — fixed by protocol where possible, residual effects modeled with ComBat or linear mixed models (site as random effect). A traveling phantom (one subject or a physical phantom scanned at all sites) quantifies residual bias.\n- **QC metrics, scored per scan** — SNR, CNR, motion (e.g., FD for fMRI, Euler number for FreeSurfer T1 QC), ghosting, coverage, geometric distortion. Define pre-specified exclusion thresholds before the analysis starts, not after.\n- **Central vs local processing** — central processing removes local pipeline variance; local processing with a harmonized Docker\u002FSingularity container is acceptable if the container is version-pinned.\n- **Reader variability** — for any human-in-the-loop step (segmentation, lesion counting, BI-RADS), use ≥ 2 readers on a sample, report Cohen's κ or ICC, and adjudicate disagreements with a predefined rule.\n- **Analysis plan** — pre-register the statistical model. Decide in advance whether site is a fixed effect (few sites, stable) or random effect (many sites, generalization target), and whether ComBat is applied at feature level or image level.\n\n### 7. Decision framework — apply to every imaging study\n\nBefore writing a single line of pipeline code, answer four questions:\n\n1. **What is the clinical question?** — Screening, diagnosis, staging, prognosis, treatment response, mechanism? The question determines the endpoint and the acceptable error profile.\n2. **What modality and contrast answers it?** — Match biology to physics (see §4). If multiple modalities answer it, pick the one with the best cost\u002Fbenefit and existing validation for the indication.\n3. **What preprocessing preserves the signal of interest?** — Work backward from the measurement. If the measurement is volume change, preserve individual anatomy. If it is group-level activation, register to a template. If it is texture, standardize acquisition and avoid interpolation across the ROI.\n4. **What confounds must be controlled?** — Age, sex, scanner, site, coil, software version, motion, time of day, medication, scan quality. List them explicitly, decide which are in the model, which are excluded by design, and which are acknowledged as limitations.\n\nIf any of the four is unclear or contested, resolve it before building the pipeline. Most imaging study failures are failures of step 1 or 3, not failures of the tool.\n\n## When NOT to Use This Skill\n\n- Selecting imaging protocols for individual patient diagnosis (clinical radiology)\n- When regulatory submission requires qualified imaging CRO oversight\n- Real-time image interpretation or radiological reporting\n\n## When to Escalate to a Human Expert\n\n- Multi-site harmonization decisions affecting primary endpoints\n- When imaging biomarker will serve as surrogate endpoint in a trial\n- Scanner-specific protocol optimization requiring vendor expertise\n\n## Common Mistakes\n\n- **Wrong:** Registering each longitudinal timepoint independently to MNI\n  **Right:** Build a within-subject template or rigidly register follow-ups to baseline first; only then normalize to MNI if needed\n  **Why:** Subject-specific atrophy is absorbed into the warp field and the longitudinal change signal is lost\n\n- **Wrong:** Relying solely on tag-based DICOM de-identification\n  **Right:** Include pixel inspection, private-tag policy, SR\u002FPDF handling, UID remapping, and defacing in the de-identification plan\n  **Why:** Burned-in PHI in pixels, private tags, SR\u002FPDF narratives, and date-encoded UIDs all survive a naive tag-level scrub\n\n- **Wrong:** Building a radiomics model using features with test-retest ICC \u003C 0.75\n  **Right:** Filter features by ICC ≥ 0.75 (preferably ≥ 0.9) and by segmentation robustness before any supervised selection\n  **Why:** Unstable features poison reproducibility — signatures built on them will not generalize to external cohorts\n\n- **Wrong:** Ignoring scanner\u002Fsite effects in multi-site imaging studies\n  **Right:** Standardize protocols, apply ComBat or mixed-effects modeling, include site as a covariate, and pre-register the analysis plan\n  **Why:** Site confounding can fully explain apparent biological effects, producing false-positive findings\n\n- **Wrong:** Fitting dozens of radiomics features to a dataset with few outcome events\n  **Right:** Enforce EPV ≥ 10 (events per variable), use nested cross-validation, and report external validation\n  **Why:** Over-fitting produces signatures that appear strong internally but fail completely on any external cohort\n\n- **Wrong:** Choosing an imaging biomarker by convenience rather than biological relevance\n  **Right:** Run the Decision Framework (question → modality → biomarker) — the biomarker should be mechanistically linked to the clinical question\n  **Why:** Using whole-brain volume when the question is hippocampal atrophy, or ADC when the question is perfusion, measures the wrong thing\n\n- **Wrong:** Deciding QC exclusion thresholds after seeing the analysis results\n  **Right:** Pre-specify QC metrics and exclusion thresholds before analysis; report inclusion\u002Fexclusion counts transparently\n  **Why:** Post-hoc threshold selection biases the study toward desired results\n\n- **Wrong:** Smoothing or interpolating through a lesion or tumor ROI\n  **Right:** Keep ROI-based measurements in native space, or use lesion-filling \u002F cost-function masking before template normalization\n  **Why:** Smoothing across lesion boundaries corrupts texture features and blurs lesion edges, invalidating radiomics and lesion-aware analyses\n\n## References\n- IBSI: Zwanenburg et al. Radiology 2020, https:\u002F\u002Fdoi.org\u002F10.1148\u002Fradiol.2020191145\n- ADNI protocol: https:\u002F\u002Fadni.loni.usc.edu\u002Fmethods\u002Fmri-tool\u002Fmri-analysis\u002F\n- ComBat harmonization: Fortin et al. NeuroImage 2018, https:\u002F\u002Fdoi.org\u002F10.1016\u002Fj.neuroimage.2017.11.024\n",{"data":50,"body":60},{"name":4,"description":6,"usage":51,"version":52,"tags":53},"Invoke when planning an imaging study, choosing preprocessing strategies, selecting imaging biomarkers, or designing radiomics pipelines.","1.0.0",[54,55,41,56,57,58,59],"skill","category:reasoning","study-design","biomarker","radiomics","hcls",{"type":61,"children":62},"root",[63,72,79,85,98,104,109,139,144,150,178,184,191,196,337,343,348,391,396,402,407,495,500,506,511,749,754,760,765,845,851,856,919,925,930,974,979,985,1003,1009,1027,1033,1207,1213],{"type":64,"tag":65,"props":66,"children":68},"element","h1",{"id":67},"imaging-study-design-and-biomarker-selection",[69],{"type":70,"value":71},"text","Imaging Study Design and Biomarker Selection",{"type":64,"tag":73,"props":74,"children":76},"h2",{"id":75},"overview",[77],{"type":70,"value":78},"Overview",{"type":64,"tag":80,"props":81,"children":82},"p",{},[83],{"type":70,"value":84},"This skill encodes methodology for designing medical imaging studies and selecting imaging biomarkers. It guides decisions about preprocessing, registration targets, de-identification, biomarker choice, radiomics stability, and multi-site harmonization. It is a reasoning skill — it does not generate pipeline code. For concrete tooling (FreeSurfer, FSL, ANTs, PyRadiomics, nnU-Net, MONAI), pair this skill with a pipeline skill after the study design is settled.",{"type":64,"tag":80,"props":86,"children":87},{},[88,90,96],{"type":70,"value":89},"The central principle: ",{"type":64,"tag":91,"props":92,"children":93},"strong",{},[94],{"type":70,"value":95},"preprocessing and biomarker choice must preserve the signal that answers the clinical question",{"type":70,"value":97},". A pipeline that is standard for one question (e.g., MNI registration for group comparison) can destroy the signal for another (e.g., individual atrophy in longitudinal studies).",{"type":64,"tag":73,"props":99,"children":101},{"id":100},"usage",[102],{"type":70,"value":103},"Usage",{"type":64,"tag":80,"props":105,"children":106},{},[107],{"type":70,"value":108},"Invoke this skill when planning an imaging study or critiquing an existing imaging pipeline. Typical prompts:",{"type":64,"tag":110,"props":111,"children":112},"ul",{},[113,119,124,129,134],{"type":64,"tag":114,"props":115,"children":116},"li",{},[117],{"type":70,"value":118},"\"I want to measure hippocampal atrophy over 2 years in Alzheimer's patients — how should I preprocess?\"",{"type":64,"tag":114,"props":120,"children":121},{},[122],{"type":70,"value":123},"\"Design a multi-site radiomics study for NSCLC prognosis.\"",{"type":64,"tag":114,"props":125,"children":126},{},[127],{"type":70,"value":128},"\"We're collecting brain MRI across three scanners. What harmonization do I need?\"",{"type":64,"tag":114,"props":130,"children":131},{},[132],{"type":70,"value":133},"\"Is it safe to share these DICOMs after standard de-identification?\"",{"type":64,"tag":114,"props":135,"children":136},{},[137],{"type":70,"value":138},"\"Which imaging biomarker should I use for presurgical mapping in epilepsy?\"",{"type":64,"tag":80,"props":140,"children":141},{},[142],{"type":70,"value":143},"Use the Decision Framework at the end of Core Concepts to structure any imaging study question.",{"type":64,"tag":73,"props":145,"children":147},{"id":146},"response-format",[148],{"type":70,"value":149},"Response Format",{"type":64,"tag":110,"props":151,"children":152},{},[153,158,163,168,173],{"type":64,"tag":114,"props":154,"children":155},{},[156],{"type":70,"value":157},"Lead with the direct recommendation or classification (≤3 sentences)",{"type":64,"tag":114,"props":159,"children":160},{},[161],{"type":70,"value":162},"Structure as: recommendation → justification (citing specific criteria\u002Fthresholds) → caveats",{"type":64,"tag":114,"props":164,"children":165},{},[166],{"type":70,"value":167},"Use tables for comparisons; bullet points for criteria lists",{"type":64,"tag":114,"props":169,"children":170},{},[171],{"type":70,"value":172},"Omit background the user already knows — they asked the question",{"type":64,"tag":114,"props":174,"children":175},{},[176],{"type":70,"value":177},"Target: 200-400 words unless the user requests exhaustive detail\nThe decision trees and frameworks in this skill are for internal reasoning only. Apply them to reach your conclusion, but do not reproduce them in your response. Present only the final recommendation with supporting evidence.",{"type":64,"tag":73,"props":179,"children":181},{"id":180},"core-concepts",[182],{"type":70,"value":183},"Core Concepts",{"type":64,"tag":185,"props":186,"children":188},"h3",{"id":187},"_1-preprocessing-strategy-is-downstream-goal-dependent",[189],{"type":70,"value":190},"1. Preprocessing strategy is downstream-goal-dependent",{"type":64,"tag":80,"props":192,"children":193},{},[194],{"type":70,"value":195},"There is no universal \"preprocess MRI\" recipe. The pipeline must match the analysis target.",{"type":64,"tag":197,"props":198,"children":199},"table",{},[200,224],{"type":64,"tag":201,"props":202,"children":203},"thead",{},[204],{"type":64,"tag":205,"props":206,"children":207},"tr",{},[208,214,219],{"type":64,"tag":209,"props":210,"children":211},"th",{},[212],{"type":70,"value":213},"Downstream goal",{"type":64,"tag":209,"props":215,"children":216},{},[217],{"type":70,"value":218},"Preprocessing priorities",{"type":64,"tag":209,"props":220,"children":221},{},[222],{"type":70,"value":223},"What to avoid",{"type":64,"tag":225,"props":226,"children":227},"tbody",{},[228,247,265,283,301,319],{"type":64,"tag":205,"props":229,"children":230},{},[231,237,242],{"type":64,"tag":232,"props":233,"children":234},"td",{},[235],{"type":70,"value":236},"Group comparison (cross-sectional)",{"type":64,"tag":232,"props":238,"children":239},{},[240],{"type":70,"value":241},"Register to common template (MNI\u002FSPM), smooth, modulate if VBM",{"type":64,"tag":232,"props":243,"children":244},{},[245],{"type":70,"value":246},"Native-space-only analysis",{"type":64,"tag":205,"props":248,"children":249},{},[250,255,260],{"type":64,"tag":232,"props":251,"children":252},{},[253],{"type":70,"value":254},"Longitudinal atrophy \u002F change",{"type":64,"tag":232,"props":256,"children":257},{},[258],{"type":70,"value":259},"Rigid register follow-ups to subject baseline, then (optionally) baseline to template",{"type":64,"tag":232,"props":261,"children":262},{},[263],{"type":70,"value":264},"Registering each timepoint independently to MNI",{"type":64,"tag":205,"props":266,"children":267},{},[268,273,278],{"type":64,"tag":232,"props":269,"children":270},{},[271],{"type":70,"value":272},"Lesion segmentation",{"type":64,"tag":232,"props":274,"children":275},{},[276],{"type":70,"value":277},"Bias field correction, skull strip, intensity normalization; registration often optional",{"type":64,"tag":232,"props":279,"children":280},{},[281],{"type":70,"value":282},"Aggressive smoothing that blurs lesion edges",{"type":64,"tag":205,"props":284,"children":285},{},[286,291,296],{"type":64,"tag":232,"props":287,"children":288},{},[289],{"type":70,"value":290},"Radiomics \u002F texture",{"type":64,"tag":232,"props":292,"children":293},{},[294],{"type":70,"value":295},"Standardize acquisition first (voxel size, reconstruction, binning); resample to isotropic; discretize intensity deterministically",{"type":64,"tag":232,"props":297,"children":298},{},[299],{"type":70,"value":300},"Mixing protocols without harmonization; default-bin radiomics on heterogeneous data",{"type":64,"tag":205,"props":302,"children":303},{},[304,309,314],{"type":64,"tag":232,"props":305,"children":306},{},[307],{"type":70,"value":308},"Functional MRI",{"type":64,"tag":232,"props":310,"children":311},{},[312],{"type":70,"value":313},"Slice-timing (if long TR), motion correction, distortion correction, coregister to anatomical, normalize if group, bandpass filter as appropriate",{"type":64,"tag":232,"props":315,"children":316},{},[317],{"type":70,"value":318},"Normalizing before motion scrubbing; smoothing before ICA denoising",{"type":64,"tag":205,"props":320,"children":321},{},[322,327,332],{"type":64,"tag":232,"props":323,"children":324},{},[325],{"type":70,"value":326},"Diffusion MRI",{"type":64,"tag":232,"props":328,"children":329},{},[330],{"type":70,"value":331},"Denoise, Gibbs unringing, eddy + motion correction, bias correction, then tensor\u002FODF fitting",{"type":64,"tag":232,"props":333,"children":334},{},[335],{"type":70,"value":336},"Reordering these steps; skipping distortion correction when reverse-PE is available",{"type":64,"tag":185,"props":338,"children":340},{"id":339},"_2-registration-target-decisions",[341],{"type":70,"value":342},"2. Registration target decisions",{"type":64,"tag":80,"props":344,"children":345},{},[346],{"type":70,"value":347},"The registration target encodes an assumption about where the signal lives.",{"type":64,"tag":110,"props":349,"children":350},{},[351,361,371,381],{"type":64,"tag":114,"props":352,"children":353},{},[354,359],{"type":64,"tag":91,"props":355,"children":356},{},[357],{"type":70,"value":358},"MNI \u002F population template",{"type":70,"value":360}," — for cross-sectional group statistics, voxelwise analyses (VBM, TBSS), atlas-based ROI extraction. Accepts that individual anatomy is partly warped away.",{"type":64,"tag":114,"props":362,"children":363},{},[364,369],{"type":64,"tag":91,"props":365,"children":366},{},[367],{"type":70,"value":368},"Subject baseline (within-subject template)",{"type":70,"value":370}," — for longitudinal analyses of change (atrophy, tumor growth, lesion evolution). Preserves each subject's anatomy and measures change in a consistent individual frame.",{"type":64,"tag":114,"props":372,"children":373},{},[374,379],{"type":64,"tag":91,"props":375,"children":376},{},[377],{"type":70,"value":378},"Native space",{"type":70,"value":380}," — for lesion analysis, surgical planning, radiomics on tumors, any case where registration would deform the structure of interest or introduce interpolation artifacts in the ROI.",{"type":64,"tag":114,"props":382,"children":383},{},[384,389],{"type":64,"tag":91,"props":385,"children":386},{},[387],{"type":70,"value":388},"Subject anatomical (T1)",{"type":70,"value":390}," — for within-subject cross-modal alignment (fMRI, DWI, PET to T1) before any group normalization.",{"type":64,"tag":80,"props":392,"children":393},{},[394],{"type":70,"value":395},"Heuristic: ask \"if I warp this image, do I still measure the thing I care about?\" If the answer is no (volume change, lesion shape, tumor texture), keep the measurement in native or within-subject space.",{"type":64,"tag":185,"props":397,"children":399},{"id":398},"_3-dicom-de-identification-risk-surface",[400],{"type":70,"value":401},"3. DICOM de-identification risk surface",{"type":64,"tag":80,"props":403,"children":404},{},[405],{"type":70,"value":406},"Standard tag-based de-identification (DICOM PS 3.15 Basic Application Level Confidentiality Profile) is necessary but not sufficient. Residual PHI risks:",{"type":64,"tag":110,"props":408,"children":409},{},[410,420,455,465,475,485],{"type":64,"tag":114,"props":411,"children":412},{},[413,418],{"type":64,"tag":91,"props":414,"children":415},{},[416],{"type":70,"value":417},"Burned-in annotations in pixel data",{"type":70,"value":419}," — ultrasound frames, screen captures, secondary captures, dose reports, 3D reformats. Patient name\u002FMRN\u002FDOB is often rendered into the pixels. Tag-level scrubbing does not remove this. Requires visual QC or OCR-based pixel redaction.",{"type":64,"tag":114,"props":421,"children":422},{},[423,428,430,437,439,445,447,453],{"type":64,"tag":91,"props":424,"children":425},{},[426],{"type":70,"value":427},"Private tags by manufacturer",{"type":70,"value":429}," — Siemens CSA headers (",{"type":64,"tag":431,"props":432,"children":434},"code",{"className":433},[],[435],{"type":70,"value":436},"0029,xx10",{"type":70,"value":438}," \u002F ",{"type":64,"tag":431,"props":440,"children":442},{"className":441},[],[443],{"type":70,"value":444},"0029,xx20",{"type":70,"value":446},") may contain protocol and sometimes patient info; GE private blocks (e.g., ",{"type":64,"tag":431,"props":448,"children":450},{"className":449},[],[451],{"type":70,"value":452},"0043,xx",{"type":70,"value":454},") and Philips MR private tags can hold acquisition metadata that includes identifiers or free-text. Default DICOM profiles may retain or only partially scrub these. Decide per-vendor: strip all private tags unless a specific block is needed, and validate.",{"type":64,"tag":114,"props":456,"children":457},{},[458,463],{"type":64,"tag":91,"props":459,"children":460},{},[461],{"type":70,"value":462},"Structured reports (SR) and encapsulated PDFs",{"type":70,"value":464}," — narrative findings often contain patient name, referring physician, dates, prior history. Must be treated as free-text and redacted or dropped.",{"type":64,"tag":114,"props":466,"children":467},{},[468,473],{"type":64,"tag":91,"props":469,"children":470},{},[471],{"type":70,"value":472},"UIDs encoding dates or MRNs",{"type":70,"value":474}," — some sites generate Study\u002FSeries\u002FSOP Instance UIDs from timestamps or record numbers. Rewrite UIDs with a consistent hash (preserving referential integrity within the release) rather than leaving originals.",{"type":64,"tag":114,"props":476,"children":477},{},[478,483],{"type":64,"tag":91,"props":479,"children":480},{},[481],{"type":70,"value":482},"Facial reconstructability from head MRI\u002FCT",{"type":70,"value":484}," — high-resolution T1\u002FCT allow face re-rendering. Apply defacing (e.g., remove or deform voxels around the face) for public release; consider skull-stripping for research-only shares.",{"type":64,"tag":114,"props":486,"children":487},{},[488,493],{"type":64,"tag":91,"props":489,"children":490},{},[491],{"type":70,"value":492},"Acquisition dates and times",{"type":70,"value":494}," — even after name removal, exact dates can re-identify when combined with auxiliary data. Date-shift per subject with a preserved offset if relative timing matters.",{"type":64,"tag":80,"props":496,"children":497},{},[498],{"type":70,"value":499},"A defensible de-identification plan specifies: which profile is applied, how private tags are handled, whether pixel data is inspected, how SR\u002FPDF objects are handled, UID remapping strategy, defacing policy, and a QC sample checked by a human.",{"type":64,"tag":185,"props":501,"children":503},{"id":502},"_4-imaging-biomarker-selection-by-indication",[504],{"type":70,"value":505},"4. Imaging biomarker selection by indication",{"type":64,"tag":80,"props":507,"children":508},{},[509],{"type":70,"value":510},"Match the biomarker to the biology being probed.",{"type":64,"tag":197,"props":512,"children":513},{},[514,540],{"type":64,"tag":201,"props":515,"children":516},{},[517],{"type":64,"tag":205,"props":518,"children":519},{},[520,525,530,535],{"type":64,"tag":209,"props":521,"children":522},{},[523],{"type":70,"value":524},"Indication",{"type":64,"tag":209,"props":526,"children":527},{},[528],{"type":70,"value":529},"Modality \u002F contrast",{"type":64,"tag":209,"props":531,"children":532},{},[533],{"type":70,"value":534},"Biomarker",{"type":64,"tag":209,"props":536,"children":537},{},[538],{"type":70,"value":539},"Rationale",{"type":64,"tag":225,"props":541,"children":542},{},[543,566,589,611,634,657,680,703,726],{"type":64,"tag":205,"props":544,"children":545},{},[546,551,556,561],{"type":64,"tag":232,"props":547,"children":548},{},[549],{"type":70,"value":550},"Alzheimer's disease, MCI",{"type":64,"tag":232,"props":552,"children":553},{},[554],{"type":70,"value":555},"Structural T1 MRI",{"type":64,"tag":232,"props":557,"children":558},{},[559],{"type":70,"value":560},"Hippocampal volume, cortical thickness, ventricular volume",{"type":64,"tag":232,"props":562,"children":563},{},[564],{"type":70,"value":565},"Measures neurodegeneration directly",{"type":64,"tag":205,"props":567,"children":568},{},[569,574,579,584],{"type":64,"tag":232,"props":570,"children":571},{},[572],{"type":70,"value":573},"Multiple sclerosis",{"type":64,"tag":232,"props":575,"children":576},{},[577],{"type":70,"value":578},"T2\u002FFLAIR + T1",{"type":64,"tag":232,"props":580,"children":581},{},[582],{"type":70,"value":583},"Lesion load, lesion count, brain parenchymal fraction, cord area",{"type":64,"tag":232,"props":585,"children":586},{},[587],{"type":70,"value":588},"Disease burden and atrophy",{"type":64,"tag":205,"props":590,"children":591},{},[592,597,601,606],{"type":64,"tag":232,"props":593,"children":594},{},[595],{"type":70,"value":596},"White matter disease, DAI, schizophrenia",{"type":64,"tag":232,"props":598,"children":599},{},[600],{"type":70,"value":326},{"type":64,"tag":232,"props":602,"children":603},{},[604],{"type":70,"value":605},"FA, MD, RD, AD (tract-based or voxelwise)",{"type":64,"tag":232,"props":607,"children":608},{},[609],{"type":70,"value":610},"Microstructural integrity",{"type":64,"tag":205,"props":612,"children":613},{},[614,619,624,629],{"type":64,"tag":232,"props":615,"children":616},{},[617],{"type":70,"value":618},"Acute stroke",{"type":64,"tag":232,"props":620,"children":621},{},[622],{"type":70,"value":623},"DWI + PWI",{"type":64,"tag":232,"props":625,"children":626},{},[627],{"type":70,"value":628},"DWI lesion volume, PWI-DWI mismatch, CBF",{"type":64,"tag":232,"props":630,"children":631},{},[632],{"type":70,"value":633},"Core vs penumbra triage",{"type":64,"tag":205,"props":635,"children":636},{},[637,642,647,652],{"type":64,"tag":232,"props":638,"children":639},{},[640],{"type":70,"value":641},"Brain tumor (glioma)",{"type":64,"tag":232,"props":643,"children":644},{},[645],{"type":70,"value":646},"DSC\u002FDCE perfusion, MRS, diffusion",{"type":64,"tag":232,"props":648,"children":649},{},[650],{"type":70,"value":651},"rCBV, Ktrans, ADC, NAA\u002FCho ratio",{"type":64,"tag":232,"props":653,"children":654},{},[655],{"type":70,"value":656},"Grade, treatment response, pseudoprogression vs recurrence",{"type":64,"tag":205,"props":658,"children":659},{},[660,665,670,675],{"type":64,"tag":232,"props":661,"children":662},{},[663],{"type":70,"value":664},"Presurgical planning (epilepsy, tumor)",{"type":64,"tag":232,"props":666,"children":667},{},[668],{"type":70,"value":669},"Task and resting-state fMRI, DTI tractography",{"type":64,"tag":232,"props":671,"children":672},{},[673],{"type":70,"value":674},"Language\u002Fmotor activation maps, tract-to-lesion distance",{"type":64,"tag":232,"props":676,"children":677},{},[678],{"type":70,"value":679},"Functional localization",{"type":64,"tag":205,"props":681,"children":682},{},[683,688,693,698],{"type":64,"tag":232,"props":684,"children":685},{},[686],{"type":70,"value":687},"Neuronal integrity (various)",{"type":64,"tag":232,"props":689,"children":690},{},[691],{"type":70,"value":692},"1H-MRS",{"type":64,"tag":232,"props":694,"children":695},{},[696],{"type":70,"value":697},"NAA\u002FCr, Cho\u002FCr, mI\u002FCr",{"type":64,"tag":232,"props":699,"children":700},{},[701],{"type":70,"value":702},"Metabolic markers of neuronal loss, membrane turnover",{"type":64,"tag":205,"props":704,"children":705},{},[706,711,716,721],{"type":64,"tag":232,"props":707,"children":708},{},[709],{"type":70,"value":710},"Oncology outside brain",{"type":64,"tag":232,"props":712,"children":713},{},[714],{"type":70,"value":715},"CT, PET\u002FCT, mpMRI",{"type":64,"tag":232,"props":717,"children":718},{},[719],{"type":70,"value":720},"Volumetry, SUVmax\u002FSUVpeak, radiomic signatures, ADC",{"type":64,"tag":232,"props":722,"children":723},{},[724],{"type":70,"value":725},"Staging, response (RECIST, PERCIST)",{"type":64,"tag":205,"props":727,"children":728},{},[729,734,739,744],{"type":64,"tag":232,"props":730,"children":731},{},[732],{"type":70,"value":733},"Cardiac",{"type":64,"tag":232,"props":735,"children":736},{},[737],{"type":70,"value":738},"Cine, LGE, T1\u002FT2 mapping",{"type":64,"tag":232,"props":740,"children":741},{},[742],{"type":70,"value":743},"EF, strain, LGE burden, native T1, ECV",{"type":64,"tag":232,"props":745,"children":746},{},[747],{"type":70,"value":748},"Function, fibrosis, edema",{"type":64,"tag":80,"props":750,"children":751},{},[752],{"type":70,"value":753},"Selection criteria: (1) biological plausibility — is this biomarker mechanistically linked to the question? (2) measurability — does the available scanner\u002Fprotocol yield adequate SNR and reliability? (3) validity — has it been validated for this indication and this population? (4) practicality — can it be obtained within the study's scan time and cost?",{"type":64,"tag":185,"props":755,"children":757},{"id":756},"_5-radiomics-stability-and-sample-size-discipline",[758],{"type":70,"value":759},"5. Radiomics stability and sample-size discipline",{"type":64,"tag":80,"props":761,"children":762},{},[763],{"type":70,"value":764},"Radiomics features are notoriously unstable. A signature built on unstable features will not generalize.",{"type":64,"tag":110,"props":766,"children":767},{},[768,778,788,798,808,818,835],{"type":64,"tag":114,"props":769,"children":770},{},[771,776],{"type":64,"tag":91,"props":772,"children":773},{},[774],{"type":70,"value":775},"Test-retest reliability",{"type":70,"value":777}," — require ICC ≥ 0.75 (commonly ICC ≥ 0.9 for high confidence) on repeat scans for any feature entering the model. Use a test-retest cohort or a publicly available one (e.g., RIDER Lung CT) aligned to the modality.",{"type":64,"tag":114,"props":779,"children":780},{},[781,786],{"type":64,"tag":91,"props":782,"children":783},{},[784],{"type":70,"value":785},"IBSI compliance",{"type":70,"value":787}," — use an IBSI-compliant extractor (PyRadiomics with IBSI settings, CERR, LIFEx) and report IBSI feature definitions and hash. Non-compliant extractors produce features that cannot be reproduced across sites.",{"type":64,"tag":114,"props":789,"children":790},{},[791,796],{"type":64,"tag":91,"props":792,"children":793},{},[794],{"type":70,"value":795},"Segmentation robustness",{"type":70,"value":797}," — evaluate inter-rater and intra-rater ICC for feature values under realistic contour variability, not just ideal contours.",{"type":64,"tag":114,"props":799,"children":800},{},[801,806],{"type":64,"tag":91,"props":802,"children":803},{},[804],{"type":70,"value":805},"Acquisition standardization first",{"type":70,"value":807}," — fixing voxel size, reconstruction kernel, and intensity discretization (fixed bin width vs fixed bin count — choose one and justify) eliminates more variance than any post-hoc harmonization.",{"type":64,"tag":114,"props":809,"children":810},{},[811,816],{"type":64,"tag":91,"props":812,"children":813},{},[814],{"type":70,"value":815},"Harmonization",{"type":70,"value":817}," — apply ComBat (or its variants: parametric\u002Fnon-parametric, with or without covariate preservation) to remove scanner\u002Fsite effects after acquisition standardization, not as a substitute for it.",{"type":64,"tag":114,"props":819,"children":820},{},[821,826,828,833],{"type":64,"tag":91,"props":822,"children":823},{},[824],{"type":70,"value":825},"Feature reduction before modeling",{"type":70,"value":827}," — cluster redundant features (|ρ| > 0.8), drop unstable features, then apply a supervised reducer (LASSO, mRMR) inside cross-validation. Cap the final model at roughly ",{"type":64,"tag":91,"props":829,"children":830},{},[831],{"type":70,"value":832},"1 feature per 10 events",{"type":70,"value":834}," (EPV ≥ 10) to avoid overfitting. For low-event studies, report this constraint explicitly; it often means choosing 3–5 features, not 20.",{"type":64,"tag":114,"props":836,"children":837},{},[838,843],{"type":64,"tag":91,"props":839,"children":840},{},[841],{"type":70,"value":842},"Report a TRIPOD- or CLEAR-style pipeline description",{"type":70,"value":844}," so the study is reproducible.",{"type":64,"tag":185,"props":846,"children":848},{"id":847},"_6-multi-site-study-considerations",[849],{"type":70,"value":850},"6. Multi-site study considerations",{"type":64,"tag":80,"props":852,"children":853},{},[854],{"type":70,"value":855},"Multi-site studies trade statistical power for between-site heterogeneity. Plan for the heterogeneity.",{"type":64,"tag":110,"props":857,"children":858},{},[859,869,879,889,899,909],{"type":64,"tag":114,"props":860,"children":861},{},[862,867],{"type":64,"tag":91,"props":863,"children":864},{},[865],{"type":70,"value":866},"Protocol standardization",{"type":70,"value":868}," — adopt ADNI-style or equivalent harmonized protocols (sequence parameters, resolution, coverage, reconstruction). Agree on inclusion criteria for scanner models and software versions.",{"type":64,"tag":114,"props":870,"children":871},{},[872,877],{"type":64,"tag":91,"props":873,"children":874},{},[875],{"type":70,"value":876},"Scanner and coil effects",{"type":70,"value":878}," — fixed by protocol where possible, residual effects modeled with ComBat or linear mixed models (site as random effect). A traveling phantom (one subject or a physical phantom scanned at all sites) quantifies residual bias.",{"type":64,"tag":114,"props":880,"children":881},{},[882,887],{"type":64,"tag":91,"props":883,"children":884},{},[885],{"type":70,"value":886},"QC metrics, scored per scan",{"type":70,"value":888}," — SNR, CNR, motion (e.g., FD for fMRI, Euler number for FreeSurfer T1 QC), ghosting, coverage, geometric distortion. Define pre-specified exclusion thresholds before the analysis starts, not after.",{"type":64,"tag":114,"props":890,"children":891},{},[892,897],{"type":64,"tag":91,"props":893,"children":894},{},[895],{"type":70,"value":896},"Central vs local processing",{"type":70,"value":898}," — central processing removes local pipeline variance; local processing with a harmonized Docker\u002FSingularity container is acceptable if the container is version-pinned.",{"type":64,"tag":114,"props":900,"children":901},{},[902,907],{"type":64,"tag":91,"props":903,"children":904},{},[905],{"type":70,"value":906},"Reader variability",{"type":70,"value":908}," — for any human-in-the-loop step (segmentation, lesion counting, BI-RADS), use ≥ 2 readers on a sample, report Cohen's κ or ICC, and adjudicate disagreements with a predefined rule.",{"type":64,"tag":114,"props":910,"children":911},{},[912,917],{"type":64,"tag":91,"props":913,"children":914},{},[915],{"type":70,"value":916},"Analysis plan",{"type":70,"value":918}," — pre-register the statistical model. Decide in advance whether site is a fixed effect (few sites, stable) or random effect (many sites, generalization target), and whether ComBat is applied at feature level or image level.",{"type":64,"tag":185,"props":920,"children":922},{"id":921},"_7-decision-framework-apply-to-every-imaging-study",[923],{"type":70,"value":924},"7. Decision framework — apply to every imaging study",{"type":64,"tag":80,"props":926,"children":927},{},[928],{"type":70,"value":929},"Before writing a single line of pipeline code, answer four questions:",{"type":64,"tag":931,"props":932,"children":933},"ol",{},[934,944,954,964],{"type":64,"tag":114,"props":935,"children":936},{},[937,942],{"type":64,"tag":91,"props":938,"children":939},{},[940],{"type":70,"value":941},"What is the clinical question?",{"type":70,"value":943}," — Screening, diagnosis, staging, prognosis, treatment response, mechanism? The question determines the endpoint and the acceptable error profile.",{"type":64,"tag":114,"props":945,"children":946},{},[947,952],{"type":64,"tag":91,"props":948,"children":949},{},[950],{"type":70,"value":951},"What modality and contrast answers it?",{"type":70,"value":953}," — Match biology to physics (see §4). If multiple modalities answer it, pick the one with the best cost\u002Fbenefit and existing validation for the indication.",{"type":64,"tag":114,"props":955,"children":956},{},[957,962],{"type":64,"tag":91,"props":958,"children":959},{},[960],{"type":70,"value":961},"What preprocessing preserves the signal of interest?",{"type":70,"value":963}," — Work backward from the measurement. If the measurement is volume change, preserve individual anatomy. If it is group-level activation, register to a template. If it is texture, standardize acquisition and avoid interpolation across the ROI.",{"type":64,"tag":114,"props":965,"children":966},{},[967,972],{"type":64,"tag":91,"props":968,"children":969},{},[970],{"type":70,"value":971},"What confounds must be controlled?",{"type":70,"value":973}," — Age, sex, scanner, site, coil, software version, motion, time of day, medication, scan quality. List them explicitly, decide which are in the model, which are excluded by design, and which are acknowledged as limitations.",{"type":64,"tag":80,"props":975,"children":976},{},[977],{"type":70,"value":978},"If any of the four is unclear or contested, resolve it before building the pipeline. Most imaging study failures are failures of step 1 or 3, not failures of the tool.",{"type":64,"tag":73,"props":980,"children":982},{"id":981},"when-not-to-use-this-skill",[983],{"type":70,"value":984},"When NOT to Use This Skill",{"type":64,"tag":110,"props":986,"children":987},{},[988,993,998],{"type":64,"tag":114,"props":989,"children":990},{},[991],{"type":70,"value":992},"Selecting imaging protocols for individual patient diagnosis (clinical radiology)",{"type":64,"tag":114,"props":994,"children":995},{},[996],{"type":70,"value":997},"When regulatory submission requires qualified imaging CRO oversight",{"type":64,"tag":114,"props":999,"children":1000},{},[1001],{"type":70,"value":1002},"Real-time image interpretation or radiological reporting",{"type":64,"tag":73,"props":1004,"children":1006},{"id":1005},"when-to-escalate-to-a-human-expert",[1007],{"type":70,"value":1008},"When to Escalate to a Human Expert",{"type":64,"tag":110,"props":1010,"children":1011},{},[1012,1017,1022],{"type":64,"tag":114,"props":1013,"children":1014},{},[1015],{"type":70,"value":1016},"Multi-site harmonization decisions affecting primary endpoints",{"type":64,"tag":114,"props":1018,"children":1019},{},[1020],{"type":70,"value":1021},"When imaging biomarker will serve as surrogate endpoint in a trial",{"type":64,"tag":114,"props":1023,"children":1024},{},[1025],{"type":70,"value":1026},"Scanner-specific protocol optimization requiring vendor expertise",{"type":64,"tag":73,"props":1028,"children":1030},{"id":1029},"common-mistakes",[1031],{"type":70,"value":1032},"Common Mistakes",{"type":64,"tag":110,"props":1034,"children":1035},{},[1036,1060,1081,1102,1123,1144,1165,1186],{"type":64,"tag":114,"props":1037,"children":1038},{},[1039,1044,1046,1051,1053,1058],{"type":64,"tag":91,"props":1040,"children":1041},{},[1042],{"type":70,"value":1043},"Wrong:",{"type":70,"value":1045}," Registering each longitudinal timepoint independently to MNI\n",{"type":64,"tag":91,"props":1047,"children":1048},{},[1049],{"type":70,"value":1050},"Right:",{"type":70,"value":1052}," Build a within-subject template or rigidly register follow-ups to baseline first; only then normalize to MNI if needed\n",{"type":64,"tag":91,"props":1054,"children":1055},{},[1056],{"type":70,"value":1057},"Why:",{"type":70,"value":1059}," Subject-specific atrophy is absorbed into the warp field and the longitudinal change signal is lost",{"type":64,"tag":114,"props":1061,"children":1062},{},[1063,1067,1069,1073,1075,1079],{"type":64,"tag":91,"props":1064,"children":1065},{},[1066],{"type":70,"value":1043},{"type":70,"value":1068}," Relying solely on tag-based DICOM de-identification\n",{"type":64,"tag":91,"props":1070,"children":1071},{},[1072],{"type":70,"value":1050},{"type":70,"value":1074}," Include pixel inspection, private-tag policy, SR\u002FPDF handling, UID remapping, and defacing in the de-identification plan\n",{"type":64,"tag":91,"props":1076,"children":1077},{},[1078],{"type":70,"value":1057},{"type":70,"value":1080}," Burned-in PHI in pixels, private tags, SR\u002FPDF narratives, and date-encoded UIDs all survive a naive tag-level scrub",{"type":64,"tag":114,"props":1082,"children":1083},{},[1084,1088,1090,1094,1096,1100],{"type":64,"tag":91,"props":1085,"children":1086},{},[1087],{"type":70,"value":1043},{"type":70,"value":1089}," Building a radiomics model using features with test-retest ICC \u003C 0.75\n",{"type":64,"tag":91,"props":1091,"children":1092},{},[1093],{"type":70,"value":1050},{"type":70,"value":1095}," Filter features by ICC ≥ 0.75 (preferably ≥ 0.9) and by segmentation robustness before any supervised selection\n",{"type":64,"tag":91,"props":1097,"children":1098},{},[1099],{"type":70,"value":1057},{"type":70,"value":1101}," Unstable features poison reproducibility — signatures built on them will not generalize to external cohorts",{"type":64,"tag":114,"props":1103,"children":1104},{},[1105,1109,1111,1115,1117,1121],{"type":64,"tag":91,"props":1106,"children":1107},{},[1108],{"type":70,"value":1043},{"type":70,"value":1110}," Ignoring scanner\u002Fsite effects in multi-site imaging studies\n",{"type":64,"tag":91,"props":1112,"children":1113},{},[1114],{"type":70,"value":1050},{"type":70,"value":1116}," Standardize protocols, apply ComBat or mixed-effects modeling, include site as a covariate, and pre-register the analysis plan\n",{"type":64,"tag":91,"props":1118,"children":1119},{},[1120],{"type":70,"value":1057},{"type":70,"value":1122}," Site confounding can fully explain apparent biological effects, producing false-positive findings",{"type":64,"tag":114,"props":1124,"children":1125},{},[1126,1130,1132,1136,1138,1142],{"type":64,"tag":91,"props":1127,"children":1128},{},[1129],{"type":70,"value":1043},{"type":70,"value":1131}," Fitting dozens of radiomics features to a dataset with few outcome events\n",{"type":64,"tag":91,"props":1133,"children":1134},{},[1135],{"type":70,"value":1050},{"type":70,"value":1137}," Enforce EPV ≥ 10 (events per variable), use nested cross-validation, and report external validation\n",{"type":64,"tag":91,"props":1139,"children":1140},{},[1141],{"type":70,"value":1057},{"type":70,"value":1143}," Over-fitting produces signatures that appear strong internally but fail completely on any external cohort",{"type":64,"tag":114,"props":1145,"children":1146},{},[1147,1151,1153,1157,1159,1163],{"type":64,"tag":91,"props":1148,"children":1149},{},[1150],{"type":70,"value":1043},{"type":70,"value":1152}," Choosing an imaging biomarker by convenience rather than biological relevance\n",{"type":64,"tag":91,"props":1154,"children":1155},{},[1156],{"type":70,"value":1050},{"type":70,"value":1158}," Run the Decision Framework (question → modality → biomarker) — the biomarker should be mechanistically linked to the clinical question\n",{"type":64,"tag":91,"props":1160,"children":1161},{},[1162],{"type":70,"value":1057},{"type":70,"value":1164}," Using whole-brain volume when the question is hippocampal atrophy, or ADC when the question is perfusion, measures the wrong thing",{"type":64,"tag":114,"props":1166,"children":1167},{},[1168,1172,1174,1178,1180,1184],{"type":64,"tag":91,"props":1169,"children":1170},{},[1171],{"type":70,"value":1043},{"type":70,"value":1173}," Deciding QC exclusion thresholds after seeing the analysis results\n",{"type":64,"tag":91,"props":1175,"children":1176},{},[1177],{"type":70,"value":1050},{"type":70,"value":1179}," Pre-specify QC metrics and exclusion thresholds before analysis; report inclusion\u002Fexclusion counts transparently\n",{"type":64,"tag":91,"props":1181,"children":1182},{},[1183],{"type":70,"value":1057},{"type":70,"value":1185}," Post-hoc threshold selection biases the study toward desired results",{"type":64,"tag":114,"props":1187,"children":1188},{},[1189,1193,1195,1199,1201,1205],{"type":64,"tag":91,"props":1190,"children":1191},{},[1192],{"type":70,"value":1043},{"type":70,"value":1194}," Smoothing or interpolating through a lesion or tumor ROI\n",{"type":64,"tag":91,"props":1196,"children":1197},{},[1198],{"type":70,"value":1050},{"type":70,"value":1200}," Keep ROI-based measurements in native space, or use lesion-filling \u002F cost-function masking before template normalization\n",{"type":64,"tag":91,"props":1202,"children":1203},{},[1204],{"type":70,"value":1057},{"type":70,"value":1206}," Smoothing across lesion boundaries corrupts texture features and blurs lesion edges, invalidating radiomics and lesion-aware analyses",{"type":64,"tag":73,"props":1208,"children":1210},{"id":1209},"references",[1211],{"type":70,"value":1212},"References",{"type":64,"tag":110,"props":1214,"children":1215},{},[1216,1229,1240],{"type":64,"tag":114,"props":1217,"children":1218},{},[1219,1221],{"type":70,"value":1220},"IBSI: Zwanenburg et al. Radiology 2020, ",{"type":64,"tag":1222,"props":1223,"children":1227},"a",{"href":1224,"rel":1225},"https:\u002F\u002Fdoi.org\u002F10.1148\u002Fradiol.2020191145",[1226],"nofollow",[1228],{"type":70,"value":1224},{"type":64,"tag":114,"props":1230,"children":1231},{},[1232,1234],{"type":70,"value":1233},"ADNI protocol: ",{"type":64,"tag":1222,"props":1235,"children":1238},{"href":1236,"rel":1237},"https:\u002F\u002Fadni.loni.usc.edu\u002Fmethods\u002Fmri-tool\u002Fmri-analysis\u002F",[1226],[1239],{"type":70,"value":1236},{"type":64,"tag":114,"props":1241,"children":1242},{},[1243,1245],{"type":70,"value":1244},"ComBat harmonization: Fortin et al. NeuroImage 2018, ",{"type":64,"tag":1222,"props":1246,"children":1249},{"href":1247,"rel":1248},"https:\u002F\u002Fdoi.org\u002F10.1016\u002Fj.neuroimage.2017.11.024",[1226],[1250],{"type":70,"value":1247},{"items":1252,"total":1347},[1253,1271,1284,1296,1311,1324,1337],{"slug":1254,"name":1254,"fn":1255,"description":1256,"org":1257,"tags":1258,"stars":26,"repoUrl":27,"updatedAt":1270},"aws-genai-ml-architect","design AWS GenAI and ML architectures","Reasoning skill for designing AWS GenAI and ML architectures for healthcare and life sciences workloads. Use when the user asks to choose between SageMaker and Bedrock, design a RAG system over medical literature, architect clinical NLP or medical imaging inference, plan genomics or drug discovery pipelines on AWS, address HIPAA\u002FPHI compliance in ML systems, design MLOps for regulated clinical models, or optimize cost for HCLS ML workloads. Triggers include \"AWS architecture\", \"SageMaker vs Bedrock\", \"HIPAA ML\", \"clinical RAG\", \"medical imaging inference\", \"genomics on AWS\", \"PHI training\", \"MLOps healthcare\", \"Bedrock guardrails\", \"HealthLake\", \"HCLS cloud architecture\", \"BAA compliance\", \"SageMaker endpoint\", \"Bedrock knowledge base\", \"clinical NLP on AWS\", \"FDA SaMD on AWS\".",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1259,1262,1265,1266,1267],{"name":1260,"slug":1261,"type":16},"Architecture","architecture",{"name":1263,"slug":1264,"type":16},"AWS","aws",{"name":18,"slug":19,"type":16},{"name":24,"slug":25,"type":16},{"name":1268,"slug":1269,"type":16},"LLM","llm","2026-07-12T08:38:07.975937",{"slug":1272,"name":1272,"fn":1273,"description":1274,"org":1275,"tags":1276,"stars":26,"repoUrl":27,"updatedAt":1283},"biomarker-discovery","guide biomarker discovery and validation","Reason about biomarker discovery and validation in HCLS — classifying biomarker intent, choosing feature-selection and cross-validation strategies, avoiding leakage, and planning external replication. Use when the user asks to discover, develop, or validate a biomarker; select features from high-dimensional omics or clinical data; design a validation study; choose evaluation metrics; justify sample size; combine multi-omics signals; or assess clinical utility. Triggers include \"discover a biomarker\", \"validate biomarker\", \"prognostic vs predictive\", \"feature selection\", \"LASSO vs elastic net\", \"nested cross-validation\", \"data leakage\", \"C-index\", \"time-dependent AUC\", \"decision curve analysis\", \"external validation cohort\", \"events per variable\", \"optimism-corrected\", \"multi-omics integration\", \"clinical utility of a biomarker\", \"is this biomarker ready\".",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1277,1278,1281,1282],{"name":1263,"slug":1264,"type":16},{"name":1279,"slug":1280,"type":16},"Bioinformatics","bioinformatics",{"name":24,"slug":25,"type":16},{"name":14,"slug":15,"type":16},"2026-07-12T08:37:49.295301",{"slug":1285,"name":1285,"fn":1286,"description":1287,"org":1288,"tags":1289,"stars":26,"repoUrl":27,"updatedAt":1295},"cdisc-compliance","reason about CDISC SDTM and ADaM implementation","Reason about CDISC SDTM and ADaM implementation for regulatory submissions. Use when the user asks about SDTM domain mapping, ADaM dataset design, controlled terminology versioning, define.xml completeness, FDA or PMDA submission requirements, query prioritization by clinical impact, SUPPQUAL usage, or CDISC compliance review. Triggers include \"SDTM mapping\", \"ADaM dataset\", \"CDISC compliance\", \"controlled terminology\", \"define.xml\", \"FDA submission data\", \"PMDA submission\", \"SDTM domain\", \"ADSL\", \"ADAE\", \"ADLB\", \"BDS structure\", \"SUPPQUAL\", \"RELREC\", \"value-level metadata\", \"CDISC CT\", \"regulatory submission data standards\", \"eCTD datasets\", \"SDTM 3.3\", \"ADaM 1.1\", \"query prioritization\", \"clinical data review\".\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1290,1291,1292],{"name":21,"slug":22,"type":16},{"name":24,"slug":25,"type":16},{"name":1293,"slug":1294,"type":16},"Regulatory Compliance","regulatory-compliance","2026-07-12T08:37:33.35594",{"slug":1297,"name":1297,"fn":1298,"description":1299,"org":1300,"tags":1301,"stars":26,"repoUrl":27,"updatedAt":1310},"cell-type-annotation","annotate single-cell RNA-seq clusters","Generate code to assign cell type labels to single-cell RNA-seq clusters using CellTypist, SingleR, marker-based annotation, or reference label transfer (scANVI\u002Fingest). Triggers on requests to \"annotate cell types\", \"label clusters\", \"run CellTypist\", \"SingleR annotation\", \"marker gene dotplot\", \"transfer labels from reference atlas\", \"cell identity\", \"automated annotation\", \"reference mapping\", \"scANVI label transfer\", \"canonical markers\", \"immune cell types\", \"hierarchical annotation\", \"majority voting CellTypist\", \"over-clustering annotation\".",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1302,1303,1306,1307],{"name":1279,"slug":1280,"type":16},{"name":1304,"slug":1305,"type":16},"Data Analysis","data-analysis",{"name":24,"slug":25,"type":16},{"name":1308,"slug":1309,"type":16},"RNA-seq","rna-seq","2026-07-12T08:38:05.443454",{"slug":1312,"name":1312,"fn":1313,"description":1314,"org":1315,"tags":1316,"stars":26,"repoUrl":27,"updatedAt":1323},"cheminformatics","calculate molecular properties with RDKit","Cheminformatics pipeline for small-molecule property calculation, filtering, and similarity analysis using RDKit. Use when the user asks to compute molecular descriptors, filter compounds by Lipinski or Veber rules, detect PAINS, calculate fingerprint similarity, run matched molecular pair analysis, generate ADMET descriptors, or process SMILES. Triggers include \"RDKit\", \"molecular descriptors\", \"Lipinski\", \"rule of five\", \"Veber\", \"PAINS\", \"pan-assay interference\", \"Morgan fingerprint\", \"Tanimoto\", \"fingerprint similarity\", \"matched molecular pair\", \"MMP\", \"mmpdb\", \"ADMET\", \"druglikeness\", \"SMILES\", \"cheminformatics\", \"compound filtering\", \"chemical similarity\".",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1317,1318,1321,1322],{"name":1279,"slug":1280,"type":16},{"name":1319,"slug":1320,"type":16},"Chemistry","chemistry",{"name":1304,"slug":1305,"type":16},{"name":14,"slug":15,"type":16},"2026-07-12T08:37:28.334619",{"slug":1325,"name":1325,"fn":1326,"description":1327,"org":1328,"tags":1329,"stars":26,"repoUrl":27,"updatedAt":1336},"claims-analytics","analyze and parse healthcare claims data","Pipeline skill for healthcare claims data parsing, analysis, and fraud detection. Use when the user asks to parse X12 837 or 835 claim files, manipulate ICD-10 CPT or HCPCS codes, detect billing pattern anomalies, profile providers against specialty peers, identify outlier billing behavior, validate NCCI edits programmatically, detect duplicate claims, run Benford's law analysis on charges, build claims data pipelines, or analyze E&M code distributions. Triggers include \"parse X12 837\", \"parse 835\", \"claims SQL\", \"ICD-10 manipulation\", \"CPT code analysis\", \"provider profiling\", \"billing outlier\", \"NCCI validation code\", \"duplicate claim detection\", \"Benford's law charges\", \"claims ETL\", \"E&M distribution analysis\", \"claims analytics pipeline\".\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1330,1331,1332,1335],{"name":1304,"slug":1305,"type":16},{"name":18,"slug":19,"type":16},{"name":1333,"slug":1334,"type":16},"Insurance","insurance",{"name":24,"slug":25,"type":16},"2026-07-12T08:37:34.815088",{"slug":1338,"name":1338,"fn":1339,"description":1340,"org":1341,"tags":1342,"stars":26,"repoUrl":27,"updatedAt":1346},"claims-billing-rules","analyze healthcare claims billing rules","Reasoning skill for healthcare claims billing rules and fraud detection logic. Use when the user asks about CMS billing rules, place of service codes, global surgery periods, modifier usage (25 59 76 77), NCCI edit logic, column 1 column 2 code pairs, mutually exclusive procedures, modifier indicators, fraud waste and abuse patterns, E&M upcoding, unbundling, phantom billing, impossible day detection, coding error versus fraud distinction, FWA investigation methodology, or claims audit logic. Triggers include \"CMS billing rules\", \"NCCI edits\", \"modifier 25\", \"modifier 59\", \"global surgery period\", \"upcoding\", \"unbundling\", \"phantom billing\", \"impossible day\", \"FWA\", \"fraud waste abuse\", \"coding error vs fraud\", \"claims audit\", \"billing compliance\", \"E&M level selection\".\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1343,1344,1345],{"name":18,"slug":19,"type":16},{"name":1333,"slug":1334,"type":16},{"name":1293,"slug":1294,"type":16},"2026-07-12T08:38:28.210856",40,{"items":1349,"total":1525},[1350,1369,1390,1400,1413,1426,1436,1446,1467,1482,1497,1512],{"slug":1351,"name":1351,"fn":1352,"description":1353,"org":1354,"tags":1355,"stars":1366,"repoUrl":1367,"updatedAt":1368},"agentcore-investigation","investigate Bedrock AgentCore runtime sessions","Investigate Bedrock AgentCore runtime sessions via CloudWatch Logs Insights — resolve session\u002Ftrace IDs, query OTEL spans, filter noise, build timelines. Use when debugging AgentCore agent sessions, tracing tool calls, or analyzing latency.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1356,1357,1360,1363],{"name":1263,"slug":1264,"type":16},{"name":1358,"slug":1359,"type":16},"Debugging","debugging",{"name":1361,"slug":1362,"type":16},"Logs","logs",{"name":1364,"slug":1365,"type":16},"Observability","observability",9427,"https:\u002F\u002Fgithub.com\u002Fawslabs\u002Fmcp","2026-07-12T08:37:22.601527",{"slug":1370,"name":1371,"fn":1372,"description":1373,"org":1374,"tags":1375,"stars":1366,"repoUrl":1367,"updatedAt":1389},"amazon-aurora-dsql","amazon aurora dsql","build applications with Aurora DSQL","Build with Aurora DSQL — manage schemas, execute queries, handle migrations, diagnose query plans, load data, and develop applications with a serverless, distributed SQL database. Covers IAM auth, multi-tenant patterns, MySQL-to-DSQL and PostgreSQL-to-DSQL schema conversion, FK replacement code generation, OCC retry patterns, ORM migration (Django\u002FHibernate\u002FRails), DDL operations, query plan explainability, SQL compatibility validation, and bulk data loading. Triggers on phrases like: DSQL, Aurora DSQL, create DSQL table, DSQL schema, migrate to DSQL, distributed SQL database, serverless PostgreSQL-compatible database, DSQL query plan, DSQL EXPLAIN ANALYZE, why is my DSQL query slow, DSQL foreign key, DSQL OCC retry, DSQL multi-region, load into DSQL, load CSV into DSQL, bulk load DSQL, aurora-dsql-loader.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1376,1379,1380,1383,1386],{"name":1377,"slug":1378,"type":16},"Aurora","aurora",{"name":1263,"slug":1264,"type":16},{"name":1381,"slug":1382,"type":16},"Database","database",{"name":1384,"slug":1385,"type":16},"Serverless","serverless",{"name":1387,"slug":1388,"type":16},"SQL","sql","2026-07-12T08:36:45.053393",{"slug":1391,"name":1392,"fn":1372,"description":1373,"org":1393,"tags":1394,"stars":1366,"repoUrl":1367,"updatedAt":1399},"aurora-dsql","aurora dsql",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1395,1396,1397,1398],{"name":1263,"slug":1264,"type":16},{"name":1381,"slug":1382,"type":16},{"name":1384,"slug":1385,"type":16},{"name":1387,"slug":1388,"type":16},"2026-07-12T08:36:42.694299",{"slug":1401,"name":1402,"fn":1372,"description":1373,"org":1403,"tags":1404,"stars":1366,"repoUrl":1367,"updatedAt":1412},"aws-dsql","aws dsql",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1405,1406,1407,1410,1411],{"name":1263,"slug":1264,"type":16},{"name":1381,"slug":1382,"type":16},{"name":1408,"slug":1409,"type":16},"Migration","migration",{"name":1384,"slug":1385,"type":16},{"name":1387,"slug":1388,"type":16},"2026-07-12T08:36:38.584057",{"slug":1414,"name":1415,"fn":1372,"description":1373,"org":1416,"tags":1417,"stars":1366,"repoUrl":1367,"updatedAt":1425},"distributed-postgres","distributed postgres",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1418,1419,1420,1423,1424],{"name":1263,"slug":1264,"type":16},{"name":1381,"slug":1382,"type":16},{"name":1421,"slug":1422,"type":16},"PostgreSQL","postgresql",{"name":1384,"slug":1385,"type":16},{"name":1387,"slug":1388,"type":16},"2026-07-12T08:36:46.530743",{"slug":1427,"name":1428,"fn":1372,"description":1373,"org":1429,"tags":1430,"stars":1366,"repoUrl":1367,"updatedAt":1435},"distributed-sql","distributed sql",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1431,1432,1433,1434],{"name":1263,"slug":1264,"type":16},{"name":1381,"slug":1382,"type":16},{"name":1384,"slug":1385,"type":16},{"name":1387,"slug":1388,"type":16},"2026-07-12T08:36:48.104182",{"slug":1437,"name":1437,"fn":1372,"description":1373,"org":1438,"tags":1439,"stars":1366,"repoUrl":1367,"updatedAt":1445},"dsql",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1440,1441,1442,1443,1444],{"name":1263,"slug":1264,"type":16},{"name":1381,"slug":1382,"type":16},{"name":1408,"slug":1409,"type":16},{"name":1384,"slug":1385,"type":16},{"name":1387,"slug":1388,"type":16},"2026-07-12T08:36:36.374512",{"slug":1447,"name":1447,"fn":1448,"description":1449,"org":1450,"tags":1451,"stars":1464,"repoUrl":1465,"updatedAt":1466},"cost-efficiency-analyzer","analyze cost efficiency and expenses","Analyzes cost structure, cost efficiency, and expense management from P&L data. Use when the user asks about costs, expenses, COGS, operating expenses, cost ratios, cost control, spending efficiency, margin compression from cost side, or wants to understand where money is going. Also use for \"are we spending too much\", \"cost breakdown\", \"expense analysis\", or \"how efficient are our operations\". NOT for revenue or top-line analysis.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1452,1455,1458,1461],{"name":1453,"slug":1454,"type":16},"Accounting","accounting",{"name":1456,"slug":1457,"type":16},"Analytics","analytics",{"name":1459,"slug":1460,"type":16},"Cost Optimization","cost-optimization",{"name":1462,"slug":1463,"type":16},"Finance","finance",3176,"https:\u002F\u002Fgithub.com\u002Fawslabs\u002Fagentcore-samples","2026-07-12T08:40:03.29555",{"slug":1468,"name":1468,"fn":1469,"description":1470,"org":1471,"tags":1472,"stars":1464,"repoUrl":1465,"updatedAt":1481},"executive-financial-briefing","generate executive financial briefings","Generates a concise executive-level financial briefing or summary suitable for a CEO, CFO, or board presentation. Use when the user asks for a summary, briefing, executive summary, board update, financial overview, financial health check, or \"how is the business doing\". Covers the full P&L picture in one page. Also use for \"give me the highlights\", \"what do I need to know\", or \"quick financial update\".",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1473,1474,1475,1478],{"name":1263,"slug":1264,"type":16},{"name":1462,"slug":1463,"type":16},{"name":1476,"slug":1477,"type":16},"Management","management",{"name":1479,"slug":1480,"type":16},"Reporting","reporting","2026-07-12T08:40:02.066471",{"slug":1483,"name":1483,"fn":1484,"description":1485,"org":1486,"tags":1487,"stars":1464,"repoUrl":1465,"updatedAt":1496},"multi-quarter-trend-analysis","analyze multi-quarter financial trends","Analyzes financial trends across multiple quarters by comparing P&L metrics over time. Use when the user wants to see trends, patterns, trajectories, or directional movement across 3 or more quarters. Also use for \"how are we trending\", \"show me the trend\", \"track performance over time\", \"quarter over quarter comparison across all quarters\", or any multi-period longitudinal analysis.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1488,1489,1490,1493],{"name":1456,"slug":1457,"type":16},{"name":1462,"slug":1463,"type":16},{"name":1491,"slug":1492,"type":16},"Financial Statements","financial-statements",{"name":1494,"slug":1495,"type":16},"Variance Analysis","variance-analysis","2026-07-12T08:40:00.79141",{"slug":1498,"name":1498,"fn":1499,"description":1500,"org":1501,"tags":1502,"stars":1464,"repoUrl":1465,"updatedAt":1511},"pdf","process and manipulate PDF documents","Use this skill whenever the user wants to do anything with PDF files. This includes reading or extracting text\u002Ftables from PDFs, combining or merging multiple PDFs into one, splitting PDFs apart, rotating pages, adding watermarks, creating new PDFs, filling PDF forms, encrypting\u002Fdecrypting PDFs, extracting images, and OCR on scanned PDFs to make them searchable. If the user mentions a .pdf file or asks to produce one, use this skill.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1503,1506,1509],{"name":1504,"slug":1505,"type":16},"Automation","automation",{"name":1507,"slug":1508,"type":16},"Documents","documents",{"name":1510,"slug":1498,"type":16},"PDF","2026-07-12T08:41:44.135656",{"slug":1513,"name":1513,"fn":1514,"description":1515,"org":1516,"tags":1517,"stars":1464,"repoUrl":1465,"updatedAt":1524},"quarterly-kpi-calculator","calculate quarterly financial KPIs","Calculates quarterly financial KPIs from P&L data. P&L figures can be provided directly by the user or fetched from the financial data MCP server. Use when the user wants KPI calculations such as Gross Margin %, EBITDA Margin %, Operating Expense Ratio, or Revenue Growth % QoQ. Also use for quarterly performance review, P&L analysis, or interpreting financial ratios against benchmarks.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1518,1519,1520,1521],{"name":1453,"slug":1454,"type":16},{"name":1304,"slug":1305,"type":16},{"name":1462,"slug":1463,"type":16},{"name":1522,"slug":1523,"type":16},"KPI","kpi","2026-07-12T08:39:59.54971",150]