This skill helps a clinician explain a condition or a scan report to a patient in plain words that ease fear rather than add to it. It translates alarming terms such as 'wear and tear' into what they mean and how common they are in people without symptoms, and it states openly what is uncertain. It also covers what the patient can do, which signs should prompt a review, and questions to check the message landed. It is a communication aid; the clinical content stays the clinician's, and it is not medical advice.
Convert, read and de-identify DICOM and NIfTI medical imagesAn AWS Labs skill for working with medical image files. It gives an agent ready commands and Python for turning DICOM series into NIfTI with dcm2niix, reading headers and pixels with pydicom, stripping patient details and private tags, loading NIfTI volumes with nibabel and arranging data in the BIDS folder layout. It also covers batch work grouped by series and a decision guide for choosing a de-identification approach. Its de-identification code is a starting point: check the results against your compliance rules, including any text burned into the images, before sharing them.HealthcareDICOMHealthcare
Pull clinical records out of HL7v2 and FHIR using PythonAWS's ehr-data-parsing skill gives Python code for pulling clinical data out of two healthcare formats, HL7v2 and FHIR R4. It covers patients' details, lab results, diagnoses and medicines, converting HL7v2 batches into CSV or pandas tables and flattening FHIR bundles. It also includes data quality checks and a conceptual mapping from HL7v2 to FHIR, using the python-hl7 and fhir.resources libraries. HL7v3, CDA, FHIR servers and SMART on FHIR sign-in are out of scope. Use synthetic or de-identified data when you try it.HealthcareAWSHealthcare
Work up a denied medical claim and draft the appealAWS's provider-denial-workup skill helps healthcare revenue-cycle teams handle a denied claim. It checks the inputs, classifies the denial from its CARC and RARC codes, decides whether to appeal, pend or accept, checks appeal deadlines, drafts a letter to the payer and lists the supporting documents needed. Missing facts are marked for a reviewer instead of being invented, and the draft is never submitted without a person's approval. It names cases to hand to a human expert, such as experimental-treatment or legal-exposure denials.HealthcareAWSHealthcare
De-identify clinical data with a local script before AI analysisThis skill from Aperivue's medical research collection guides a researcher through removing protected health information from CSV, TSV or Excel files before any AI-assisted analysis. The removal itself happens in a Python script on the researcher's computer that never goes online or calls a model, using patterns for 12 countries and an interactive review in the researcher's own terminal. The agent explains and checks the process but is told never to ask for raw patient data or to read the mapping file of original values. It covers assessment, scanning, review, de-identification and verification, with references for the HIPAA Safe Harbor list and date shifting. Never paste or upload raw patient data to the agent, and check the output before sharing it, since a pattern-based tool can miss identifiers.HealthcareHealthcare dataPrivacy
Assign ICD-10-CM codes to a clinical noteThis skill reads a clinical note and picks the billable ICD-10-CM diagnosis codes for a claim. It decides which diagnoses should be billed, codes each one as specifically as the note allows, and looks up the exact code through an ICD-10 connector before a final check.HealthcareMedical CodingBilling
Abstract clinical notes into cited data fieldsThis skill extracts the fields you define from clinical notes and ties each value to the exact text it came from. Missing values stay empty instead of being guessed. Output can feed chart abstraction, a patient registry, or a research cohort, and it can be exported as FHIR.HealthcareClinical NotesData Extraction
Convert, read and de-identify DICOM and NIfTI medical images
Pull clinical records out of HL7v2 and FHIR using Python
Work up a denied medical claim and draft the appeal
De-identify clinical data with a local script before AI analysis
Assign ICD-10-CM codes to a clinical note