hcls-agent-skills

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/PHI 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".

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".

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".

Generate code to assign cell type labels to single-cell RNA-seq clusters using CellTypist, SingleR, marker-based annotation, or reference label transfer (scANVI/ingest). 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".

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".

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".

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".

Reason about clinical data terminology standards — MedDRA hierarchy (LLT→PT→HLT→HLGT→SOC), ICD-10 code structure and grouping, SNOMED CT concept model, LOINC panel relationships, and mapping decisions between systems. Use when the user asks to code adverse events, map diagnoses to ICD-10, choose a coding granularity level, group AEs by SOC or PT, interpret SNOMED CT relationships, select LOINC codes for lab panels, convert between terminology systems, or decide when to aggregate at HLT vs PT level. Triggers include "MedDRA coding", "ICD-10 grouping", "SNOMED CT", "LOINC panel", "adverse event coding", "terminology mapping", "SOC table", "preferred term", "code hierarchy", "clinical coding", "AE frequency table", "diagnosis grouping", "lab code selection", "cross-walk between terminologies".

DICOM and NIfTI medical image processing pipeline. Triggers on DICOM, NIfTI, dcm2niix, de-identification, pydicom, DICOM header, conversion, anonymization, BIDS, DICOM tags, medical image format conversion, "DICOM to NIfTI", "burned-in PHI", "SeriesInstanceUID", "nibabel", "DICOM anonymization".

Generate correct code for whole-slide image (WSI) analysis using TIAToolbox and foundation models (H-optimus-0, UNI, Prov-GigaPath). Triggers on requests involving whole-slide images, WSI, digital pathology, histopathology, SVS/NDPI/pyramidal TIFF, tissue segmentation, patch extraction, stain normalization, H-optimus-0, TIAToolbox, CAMELYON16/17, SlideGraph, MIL aggregation, HoVer-Net, PanNuke, or SageMaker deployment of pathology models. Produces deterministic commands and Python snippets for slide-info inspection, tissue masking, tile extraction at specified mpp, foundation-model feature embedding, slide-level aggregation, and regulated cloud inference.

Reason about drug repurposing strategies in HCLS — choosing between target-based and phenotype-based approaches, evaluating mechanism-of-action overlap, querying drug-gene interaction databases, assessing clinical translatability, and ranking candidates by evidence strength. Use when the user asks to repurpose a drug, find approved drugs for a new indication, evaluate a repurposing candidate, query DGIdb or OpenTargets, assess drug-target interactions, design a repurposing study, rank repurposing evidence, or evaluate translatability of a candidate. Triggers include "drug repurposing", "repurpose", "repositioning", "new indication", "off-label use", "target-based repurposing", "phenotype-based repurposing", "CMap", "L1000", "DGIdb", "OpenTargets", "DrugBank", "ChEMBL", "mechanism of action overlap", "drug-gene interaction", "translatability", "existing safety data", "repurposing evidence hierarchy".

Generate code for EDC export validation, clinical data range checks, cross-form consistency checks, SDTM structure validation, controlled terminology verification, and define.xml generation. Use when the user asks to validate clinical trial data exports, check vital sign or lab value ranges, verify AE date consistency, validate SDTM datasets against CDISC rules, generate define.xml, or build an automated data review pipeline. Triggers include "EDC validation", "range check clinical data", "cross-form consistency", "SDTM validation", "controlled terminology check", "define.xml generation", "Medidata Rave export", "Oracle InForm", "Veeva Vault CDMS", "clinical data cleaning", "edit check", "data query", "SDTM structure check", "lab range check", "vital signs validation", "AE date check", "protocol deviation detection", "OpenCDISC", "Pinnacles 21", "P21 validation".

Parse and extract clinical data from HL7v2 messages and FHIR R4 resources using Python. Use when the user mentions HL7v2, HL7, FHIR, PID segment, OBX segment, MSH segment, Patient resource, Observation resource, Condition resource, MedicationRequest, EHR data extraction, clinical message parsing, FHIR bundle, HL7 to FHIR conversion, ADT message, ORU message, lab result extraction, or clinical data quality checks. Triggers include "parse HL7", "extract FHIR", "HL7v2 message", "FHIR resource", "PID segment", "OBX segment", "Patient resource", "Observation resource", "EHR parsing", "clinical data extraction", "HL7 to CSV", "FHIR to DataFrame".

Reason about germline and somatic variant classification using ACMG/AMP 2015 and AMP/ASCO/CAP frameworks. Use when the user asks to classify a variant, interpret a VCF annotation, resolve a VUS, apply ACMG criteria, weigh ClinVar evidence, evaluate gnomAD allele frequencies, interpret REVEL/CADD/SpliceAI scores, decide whether PVS1 applies, or assess gene-disease validity before reporting. Triggers include "ACMG", "variant classification", "pathogenic", "likely pathogenic", "VUS", "benign", "ClinVar", "gnomAD", "REVEL", "CADD", "SpliceAI", "PVS1", "loss of function", "nonsense variant", "missense interpretation", "splice variant", "filtering allele frequency", "ClinGen", "somatic variant tier".

Reasoning skill for HEDIS measure specification, enrollment logic, exclusion evaluation, NCQA audit requirements, and care gap prioritization. Use when the user asks about HEDIS measure definitions, denominator/numerator/exclusion logic, continuous enrollment rules, Star Rating impact, or care gap closure strategies.

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".

Reason about ML experiment design for healthcare and life sciences data. Use when the user asks to design an ML study, choose a model for clinical/biomedical data, set up cross-validation, pick evaluation metrics, audit fairness, plan a regulatory submission, or critique an ML pipeline on EHR, medical imaging, genomics, molecules, or clinical text. Triggers include "design an ML experiment", "which model for this clinical data", "how should I split", "nested CV", "class imbalance", "AUROC vs AUPRC", "calibration", "decision curve", "net benefit", "TRIPOD+AI", "PROBAST", "CLAIM", "FDA SaMD", "PCCP", "GMLP", "site generalization", "temporal leakage", "scaffold split", "foundation model evaluation", "subgroup fairness", "is this model ready for deployment".

Molecular docking pipeline using AutoDock Vina for structure-based drug discovery. Triggers on docking, AutoDock Vina, receptor preparation, ligand preparation, PDBQT, grid box, virtual screening, binding affinity, pose prediction, structure-based virtual screening, "redocking RMSD", "Vina score", "docking pose", "prepare receptor", "ligand library screening".

Reasoning skill for multi-omics data integration strategy selection. Use when the user asks to integrate transcriptomics with proteomics, combine multi-omic layers, choose between early intermediate or late integration, apply batch correction across omics, handle partial sample overlap, run MOFA+ or iCluster, interpret multi-omic factors, select enrichment methods for multi-omic signatures, or decide how to merge genomics epigenomics transcriptomics proteomics and metabolomics data. Triggers include "multi-omics integration", "combine omics layers", "early vs late fusion", "MOFA+", "iCluster", "batch correction across omics", "partial overlap", "multi-omic enrichment", "kernel integration", "concatenation vs stacking", "SNF", "similarity network fusion", "intermediate integration", "multi-omic factor analysis".

Pipeline skill for multi-omics data processing and integration. Use when the user asks to map gene IDs between HGNC Ensembl and UniProt, convert between omic data formats, run batch correction with ComBat or ComBat-seq, perform GSEA or over-representation analysis on multi-omic results, run consensus clustering on integrated data, execute MOFA2 in R or mofapy2 in Python, build a multi-omics ETL pipeline, harmonize feature identifiers across omics layers, or run clusterProfiler enrichment. Triggers include "map Ensembl to HGNC", "ID mapping omics", "ComBat code", "run MOFA2", "mofapy2", "GSEA Python", "fgsea R", "consensus clustering", "multi-omics pipeline", "gseapy", "biomaRt", "clusterProfiler", "mixOmics DIABLO".

NGS quality control pipeline for short-read sequencing data. Triggers on FastQC, QC, quality control, adapter trimming, coverage, mosdepth, Picard metrics, fastp, MultiQC, sequencing QC, BAM QC, WGS/WES coverage analysis.

Reasoning skill for prior authorization clinical policy evaluation. Use when the user asks about payer clinical criteria, step therapy requirements, medical necessity definitions, CMS LCD/NCD coverage rules, appeals documentation strategy, formulary tier implications, or FHIR Da Vinci PAS implementation guidance. Triggers include "prior auth policy", "step therapy", "medical necessity", "coverage determination", "LCD", "NCD", "formulary tier", "PA appeal", "peer-to-peer review", "Da Vinci PAS", "clinical criteria", "PA denial", "drug authorization", "utilization management".

Pipeline skill for automating prior authorization decision workflows. Use when the user asks to parse PA request data (X12 278 or FHIR PAS bundles), extract clinical features for adjudication, build rules-based PA decision engines, train ML classifiers on historical PA decisions, analyze denial patterns, or generate SHAP explanations for PA outcomes. Triggers include "parse 278", "FHIR PAS bundle", "PA automation", "adjudication logic", "PA classifier", "denial analysis", "prior auth ML", "SHAP explainability", "PA feature extraction", "rules engine PA", "PA decision pipeline", "authorization workflow", "clinical criteria extraction", "denial pattern mining", "PA turnaround time".

Reason about pharmacoepidemiologic study design for causal inference from real-world data — choosing active comparator new-user designs, emulating target trials, avoiding immortal time bias, handling time-varying confounding with marginal structural models, and selecting propensity score methods. Use when the user asks to design a drug safety or effectiveness study, choose between propensity score matching vs weighting vs stratification, emulate a target trial, handle immortal time bias, apply marginal structural models, assess unmeasured confounding with E-values, select an active comparator, define a new-user cohort, or evaluate a pharmacoepidemiology study design. Triggers include "active comparator", "new-user design", "target trial emulation", "immortal time bias", "propensity score matching", "IPTW", "marginal structural model", "confounding by indication", "E-value", "negative control outcomes", "quantitative bias analysis", "time-varying confounding", "prevalent user bias", "washout period", "landmark

Pipeline skill for protein structure analysis covering PDB/mmCIF parsing, RMSD superposition, Ramachandran/dihedral analysis, binding pocket detection, contact maps, B-factor flexibility, DSSP secondary structure, and format conversion. Triggers on PDB, protein structure, RMSD, Ramachandran, binding pocket, Biopython, PyMOL, pocket detection, fpocket, DSSP, PDBQT, structural alignment, superposition.

Pipeline skill for computing HEDIS quality measures from claims and clinical data. Use when the user asks to calculate HEDIS measure rates, check continuous enrollment, build denominator/numerator logic, detect care gaps, compute utilization rates, identify high-cost claimants, or score risk stratification indices. Triggers include "calculate HEDIS", "measure rate", "continuous enrollment check", "care gap detection", "denominator query", "numerator logic", "utilization rate", "high-cost claimant", "Charlson score", "LACE score", "claims analysis", "quality measure SQL".

Reason about quantitative proteomics experiment design and data analysis strategy. Use when the user asks to choose between LFQ, TMT, and DIA quantification; select an imputation method for missing values; pick a normalization strategy; interpret differential expression results from proteomics data; evaluate ratio compression; or design a proteomics study for biomarker discovery or validation. Triggers include "LFQ vs TMT", "DIA quantification", "proteomics normalization", "missing value imputation", "MNAR", "MinProb", "QRILC", "kNN imputation", "VSN normalization", "median centering", "quantile normalization", "limma proteomics", "ratio compression", "proteomics study design", "label-free quantification", "tandem mass tag", "data-independent acquisition", "DIA-NN", "Spectronaut", "MaxQuant LFQ", "proteomics differential expression", "empirical Bayes proteomics", "proteinGroups.txt", "MSFragger output", "TMT normalization code", "LFQ analysis", "proteomics pipeline R", "MaxQuant output".

Structural MRI/CT preprocessing pipeline for radiology workflows covering skull stripping, bias field correction, registration, and intensity normalization. Triggers on skull stripping, bias correction, registration, ANTs, FSL, HD-BET, N4, brain extraction, normalization, FLIRT, FNIRT, SyN, fslreorient2std, MNI registration, T1 preprocessing.

Pipeline skill for CMS-HCC risk adjustment calculation and coding gap identification. Use when the user asks to apply the ICD-10-to-HCC crosswalk, calculate RAF scores, resolve disease hierarchies programmatically, identify coding gaps from Rx or lab proxies, project member-level risk scores, build a risk adjustment data pipeline, compute HCC coefficients, run hierarchy resolution code, or generate member risk score reports. Triggers include "calculate RAF score", "ICD-10 to HCC crosswalk", "hierarchy resolution code", "coding gap detection", "Rx proxy gap", "lab proxy gap", "risk score projection", "HCC pipeline", "member risk report", "risk adjustment Python", "risk adjustment SQL", "RAF calculation code", "CMS-HCC pipeline".

Reasoning skill for CMS-HCC risk adjustment strategy and methodology. Use when the user asks about CMS-HCC model versions V24 or V28, blended transition methodology, ICD-10-to-HCC mapping logic, disease interaction hierarchies, RAF score methodology, risk adjustment factor calculation, coding gap identification, audit-defensible documentation, HCC recapture strategy, prospective vs retrospective risk adjustment, or Medicare Advantage risk scoring. Triggers include "CMS-HCC", "V24", "V28", "blended transition", "HCC mapping", "disease hierarchy", "RAF score", "risk adjustment", "coding gap", "HCC recapture", "chart review", "audit defensible", "risk score methodology", "Medicare Advantage risk", "capitation revenue", "hierarchical condition category".

Reasoning skill for clinical risk stratification index selection and interpretation. Use when the user asks about LACE scores, Charlson Comorbidity Index, Elixhauser Index, readmission risk scoring, comorbidity weighting, SDOH Z-codes, Area Deprivation Index, or population health stratification methods.

Bulk RNA-seq analysis pipeline covering alignment (STAR), quantification (Salmon, featureCounts), and differential expression (DESeq2). Triggers on RNA-seq, STAR, Salmon, DESeq2, differential expression, gene expression, tximport, featureCounts, transcriptomics, "bulk RNA-seq", "STAR alignment", "Salmon quantification", "gene counts", "DESeq2 results", "shrinkage estimator", "volcano plot RNA-seq".

Real-world data cohort analysis pipeline for claims and EHR data — cohort identification with ICD-10, NDC, and CPT codes, medication adherence metrics (PDC/MPR), propensity score estimation and balance diagnostics, Kaplan-Meier survival curves, and Cox proportional hazards models. Use when the user mentions claims data, cohort identification, ICD-10 codes, NDC codes, CPT codes, medication adherence, PDC, MPR, propensity score, SMD balance, Kaplan-Meier, Cox model, Schoenfeld residuals, survival analysis on claims, or real-world evidence pipeline.

Scanpy-based single-cell RNA-seq analysis pipeline covering loading (10X, h5ad), QC, normalization, HVG selection, PCA/UMAP, Leiden clustering, differential expression, and batch correction (Harmony, scVI). Use when the user mentions single-cell, scRNA-seq, Scanpy, AnnData, UMAP, clustering, 10X, h5ad, leiden, highly variable genes, Harmony, scVI, cluster cells, find marker genes, filter low-quality cells, gene expression matrix, doublet removal, normalize counts, dimensionality reduction, cell clustering, QC single-cell, filtered_feature_bc_matrix, cellranger output, 10X h5, scrublet, cell ranger, count matrix, find cell types, batch effect, integration, neighborhood graph, or differential expression genes.

Reasoning skill for structure-based drug design strategy. Use when the user asks to assess target druggability, choose a docking strategy, select a scoring function, define a binding site, interpret docking results, plan molecular dynamics or free energy perturbation (FEP), design a virtual screening cascade, or decide between fragment-based and HTS screening. Triggers include "drug design", "druggability", "docking strategy", "scoring function", "binding pocket", "molecular dynamics", "FEP", "SAR", "virtual screening", "induced fit", "allosteric site", "fragment-based drug discovery", "lead optimization", "MM-GBSA", "pharmacophore".

Single-cell trajectory inference pipeline covering diffusion pseudotime (DPT), PAGA, RNA velocity with scVelo, and fate mapping with CellRank. Use when the user mentions pseudotime, trajectory, lineage, differentiation, RNA velocity, scVelo, CellRank, PAGA, diffusion map, fate probabilities, terminal states, spliced/unspliced, velocyto, order cells by development, cell differentiation path, cell fate, branching analysis, monocle, developmental trajectory, stem cell differentiation, progenitor to mature, how do cells differentiate, cell lineage tree, velocity arrows, fate mapping, transition probabilities, root cells, terminal states, absorption probabilities, or latent time.

Reason about translational research problems in HCLS — especially neurology and hypothesis validation. Use when the user asks to design a validation study, evaluate a target, qualify a biomarker, choose endpoints, critique a preclinical-to-clinical plan, pick a trial design (basket/umbrella/platform), plan multimodal data integration, or assess whether a hypothesis is ready to advance. Triggers include phrases like "validate target", "biomarker context of use", "bench to bedside", "T0/T1/T2/T3/T4", "proof of mechanism", "target engagement", "design a trial for", "preclinical model translates", "imaging-genetics", "is this ready for Phase 2", "mouse results to humans", "GWAS/eQTL/MR", "endpoint selection", "drug repurposing rationale", "fail fast".

Germline and somatic short-variant calling pipeline for Illumina short-read data. Use when the user mentions BWA, BWA-MEM2, GATK, HaplotypeCaller, Mutect2, VCF, GVCF, VQSR, variant calling, germline, somatic, SNV, or indel calling from FASTQ/BAM.