菜单
Anonymizes clinical text by removing or replacing personally identifiable information (PHI/PII) such as names, dates, locations, identifiers, and contact details. Use when processing clinical notes, patient records, or any sensitive healthcare data.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
Classifies clinical text into predefined categories such as severity, urgency, or clinical type based on patient information. Use when assigning labels for triage, prioritization, or categorization of medical cases.
Extracts and organizes clinical events into a chronological timeline based on patient records or medical notes. Use when reconstructing the sequence of symptoms, diagnoses, treatments, or hospital events.
Converts unstructured clinical text into structured JSON by extracting key medical information such as symptoms, diagnoses, medications, procedures, and temporal data. Use when transforming clinical notes into machine-readable format.
Generates structured clinical reports (e.g., discharge summaries, medical reports, or clinical notes) based on patient data. Use when transforming clinical information into clear, standardized medical documentation.
Answers questions based strictly on clinical text, such as patient records or medical notes. Use when extracting specific information from clinical documents in a reliable and grounded manner.
Assesses patient risk level (e.g., low, medium, high) based on clinical information such as symptoms, history, and findings. Use when evaluating potential severity, complications, or need for prioritization in clinical contexts.
基于 SOC 职业分类
| name | clinical-anonymization |
| description | Anonymizes clinical text by removing or replacing personally identifiable information (PHI/PII) such as names, dates, locations, identifiers, and contact details. Use when processing clinical notes, patient records, or any sensitive healthcare data. |
When anonymizing clinical text, follow these steps:
Detect and classify all Personally Identifiable Information (PII/PHI), including:
Replace each entity with a consistent placeholder:
Rules:
If exact anonymization is not possible, generalize:
Before returning the result:
Return only the anonymized text.
Do not include explanations. Do not include metadata unless explicitly requested.
