Skill UI

This skill guides the creation of computable phenotypes and patient cohorts using the OHDSI/OMOP CDM standards. It allows users to define precise, executable criteria by combining standard concept sets (like PheKB) with advanced Natural Language Processing (NLP) features extracted from free clinical text. Use this when structured codes alone are insufficient to capture critical health information (e.g., symptoms, social context).

A powerful, on-device tool for removing, masking, or replacing sensitive Protected Health Information (PHI/PII) from clinical free text. It ensures patient privacy by running fully locally, supporting methods like masking, hashing, and date shifting. Essential for preparing medical records for safe sharing, storage, or analysis, adhering to standards like HIPAA and GDPR.

This skill de-identifies sensitive PII from clinical notes across numerous languages, including Spanish, German, French, and Arabic. It automatically selects language-specific PII models, national ID regex patterns, and generates locale-aware fake surrogates. Designed for global healthcare datasets, it runs entirely on-device, ensuring maximum privacy and linguistic accuracy for diverse international clinical records.

This skill implements a comprehensive logging guard to prevent Protected Health Information (PHI) leakage into logs, traces, and error reports. It acts as a pre-emission filter, redacting structured identifiers (SSN, email, phone) via regex and handling free-text PHI using advanced models. Essential for maintaining strict data privacy and compliance when shipping logs to centralized stores or error trackers.

This skill evaluates de-identification and clinical NER models against rigorous leakage-first release gates (G1a-G8). Instead of relying on aggregate F1 scores, it focuses on detecting residual Protected Health Information (PHI) leakage and enforcing per-label recall floors, which is crucial for HIPAA compliance. Determines if a model is RELEASABLE or QUARANTINED.

This skill facilitates the extraction of population-scale (cohort-level) clinical text from Electronic Health Records (EHR) or FHIR data warehouses using the FHIR Bulk Data Access ($export) operation. It manages the asynchronous job kickoff, polling, and downloading of NDJSON files. The resulting data stream is designed to be fed in batches into OpenMed for mandatory de-identification and Named Entity Recognition (NER) analysis.

Use openmed.extract_pii to detect and retrieve specific types of Protected Health Information (PHI) or Personally Identifiable Information (PII) spans (names, dates, MRNs, etc.) from clinical text. Unlike deidentification, this function returns the original text and the detected spans, allowing users to audit, preview redactions, normalize labels, or route spans to a custom anonymizer without altering the source data. It is ideal for pre-processing and policy enforcement.

This skill extracts crucial Social Determinants of Health (SDOH) — such as housing instability, food insecurity, and transportation barriers — from unstructured clinical notes. It maps the identified text spans to standardized ICD-10-CM Z-codes (Z55–Z65). It is essential for health-equity analytics, building risk stratification tools, and identifying coding gaps by converting narrative text into queryable, discrete data points.

Synthea generates statistically realistic, fully synthetic patient records in multiple standards (FHIR R4, C-CDA, CSV). This tool is essential for developers and researchers needing safe, shareable test data for OpenMed pipelines, CI fixtures, or specialized de-identification leakage gate test sets. It ensures zero risk of exposing real Protected Health Information (PHI).

This skill replaces detected Protected Health Information (PHI) in clinical notes with realistic, type-matched synthetic surrogates. Unlike simple redaction, surrogates maintain the natural readability and structure of the text, ensuring the data remains parseable for downstream NLP tasks, QA, or human review. It supports consistent, reproducible fake identities and allows custom identifier generation for proprietary formats.

This tool monitors the health of a Git worktree by checking background verification tasks (type checking, unit testing, and build compilation). It analyzes dedicated log files to provide a non-blocking, actionable summary of whether the branch is stable, helping developers quickly identify failures before committing.

This skill implements an advanced reference pattern for integrating with Monitoring Control Platform (MCP) servers. It ensures accurate data retrieval by first reading a domain-specific 'gotchas' reference file, mitigating query failures caused by complex server behaviors, pagination limits, or subtle API quirks. It analyzes issue data, groups errors by root cause, and generates a structured, actionable markdown report for improved system health visibility.