save-report

Save Medical Test Report Skill

Save user-provided medical test reports to the personal medical data center.

Core Process

用户输入图片 -> 读取并分析图片 -> 提取数据 -> 生成JSON -> 保存图片和数据 -> 更新索引 -> 确认输出

Step 1: Check Parameters

• image_path (required): local path of the examination report image
• exam_date (optional): examination date, format YYYY-MM-DD

Step 2: Date Determination Rules (Important)

Priority order:

1. exam_date provided by the user(highest priority)
2. Sampling time in the image
3. Sample submission time in the image
4. Testing time/report time in the image
5. Other date identifiers in the image
6. Current date (only as a fallback)

Step 3: Read and Analyze the Image

Image Analysis Prompt Templates

Biochemical Test Prompt:

请详细识别这张医疗检验报告单中的所有信息，包括：
1. 日期时间信息（采样/送检/检测/报告时间）
2. 医院/检验机构名称
3. 检验项目和结果（项目名、数值、单位、参考范围、异常标识）

Imaging Test Prompt: Use the corresponding prompt template according to the examination type:

• B-ultrasound/Color Doppler Ultrasound
• X-ray Examination
• CT Examination
• MRI Examination
• Endoscopy Examination
• Pathology Examination
• Electrocardiogram Examination
• Mammography Examination
• PET-CT Examination

Step 4: Generate the Data File

File Path Format

• Biochemical tests: data/biochemical-tests/YYYY-MM/YYYY-MM-DD_test-type.json
• Imaging studies: data/imaging-studies/YYYY-MM/YYYY-MM-DD_exam-type_body-part.json

Data Structure Definition

For the complete data structure definition, please refer to schema.json file. This file uses the JSON Schema format to define two report types:

Biochemical Test (BiochemicalTestReport)

• id: unique identifier
• type: fixed value "Biochemical Test"
• date: examination date (YYYY-MM-DD)
• hospital: name of the hospital/testing institution
• original_image: original image path
• items: array of test items
  • name: item name
  • value: test value
  • unit: unit
  • min_ref: minimum value of the reference range
  • max_ref: maximum value of the reference range
  • is_abnormal: whether abnormal

Imaging Test (ImagingTestReport)

• id: unique identifier
• type: fixed value "Imaging Test"
• subtype: examination type (B-ultrasound/CT/MRI/X-ray/Endoscopy/Pathology/Electrocardiogram/Mammography/PET-CT)
• date: examination date (YYYY-MM-DD)
• hospital: name of the hospital/testing institution
• body_part: examination site
• findings: examination results
  • description: description of examination findings
  • targets: array of target lesions (optional)
    • name: lesion name
    • location: location
    • size: size (length/width/height/unit)
    • characteristics: characteristics (shape/margin/density)
  • conclusion: examination conclusion
• original_image: original image path

Step 5: Save the Data

1. Create the month directory (if it does not exist)
2. Create the images subdirectory (if it does not exist)
3. Copy the original image to the corresponding directory
4. Save the JSON data file
5. Update the global index data/index.json

Step 6: Update the Index

In data/index.json add the index information of the new record.

Execution Instructions

1. 验证图片路径
2. 读取 schema.json 确认数据结构定义
3. 使用 mcp__4_5v_mcp__analyze_image 分析图片
4. 提取数据（优先使用用户提供的日期）
5. 根据 schema.json 生成符合规范的JSON数据文件
6. 复制图片到images目录
7. 保存JSON文件
8. 更新index.json
9. 显示确认信息

Schema Validation

The generated JSON data must conform to the structure defined in schema.json . The Schema uses the oneOf keyword to distinguish the two report types:

• Biochemical Test Report (type: "生化检查")
• Imaging Examination Report (type: "影像检查")

Example Interaction

Automatically extract date

用户: @医疗报告/血液检查.jpg

输出:
✅ 检查单保存成功！
类型：生化检查（血液常规）
日期：2025-10-07（从图片提取）
提取到 15 项检查指标
文件路径：data/生化检查/2025-10/2025-10-07_血液常规.json

Manually specify date

用户: @医疗报告/血液检查.jpg 2025-10-07

输出:
✅ 检查单保存成功！
类型：生化检查（血液常规）
日期：2025-10-07（使用用户指定日期）