← Back to Skills Marketplace
unisound-llm

unisound-overall-report

by Unisound-LLM · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ Security Clean
108
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install unisound-overall-report
Description
体检报告整体解读。输入完整体检报告文本,输出整体健康评级、主要发现(含严重程度)、优先行动清单和个性化生活方式建议(JSON + 通俗易懂的整体解读)。
README (SKILL.md)

体检报告整体解读

概述

面向体检中心/健康管理机构,给定受检者完整体检报告,本技能会:

  • 综合所有检查结果给出整体健康评级(优秀/良好/需关注/需重视)
  • 提取关键异常发现,按重要程度排序并标注严重程度
  • 列出最优先需要处理的行动(最多3条)
  • 给出个性化生活方式建议
  • 建议下次体检时间及重点关注项目
  • 生成面向受检者的通俗易懂整体解读文字

数据安全、隐私与伦理声明

  • 最小必要原则:仅处理体检报告解读所必需的检查结果数据;不要求包含姓名、证件号、手机号等直接身份标识。
  • 严格脱敏:在发送至任何模型/接口前,会对可识别个人身份的信息进行脱敏/去标识化处理。
  • 不做本地持久化:不将用户输入与中间结果写入本地持久化存储。仅在内存中短暂处理;本次调用结束即销毁
  • 医疗边界:本技能输出为健康管理辅助解读,不构成医疗诊断;如有异常项目,请务必就诊由执业医生进一步评估。

输入格式

纯文本(UTF-8),包含体检报告内容(可粘贴报告原文),例如:

受检者:男,45岁
体检日期:2026年4月

一般检查:身高175cm,体重82kg,BMI 26.8,血压142/90mmHg
血常规:WBC 7.2×10⁹/L,RBC 5.1×10¹²/L,HGB 155g/L(均正常)
生化:ALT 52U/L(↑),AST 38U/L,TG 2.8mmol/L(↑),TC 5.9mmol/L(↑),FBG 6.2mmol/L(↑)
尿常规:尿蛋白(-),尿糖(-)
心电图:窦性心律,ST段轻度改变
胸部CT:双肺未见明显实质性病变
腹部超声:轻度脂肪肝

也支持 JSON 格式(包含 text/content/report 字段的对象)。

快速开始

# 从 skills 目录运行
python3 health-exam/report-interpret/overall-report/scripts/run.py \
  --input data/health-exam-overall/case-001.txt \
  --appkey \x3Cyour-appkey>

# 保存输出到文件
python3 health-exam/report-interpret/overall-report/scripts/run.py \
  --input data/health-exam-overall/case-001.txt \
  --appkey \x3Cyour-appkey> \
  --output runs/health-exam-overall/case-001.json

参数说明

  • --input PATH必填。体检报告文件路径(txt 或 json,UTF-8)。
  • --appkey STRING必填。调用内部医疗大模型的鉴权 key,由平台分配。
  • --output PATH:输出文件路径(默认:打印到 stdout)。
  • --base URL:内部大模型 base URL(默认:https://maas-api.hivoice.cn/v1)。
  • --model STRING:模型名称(默认:u1-insuremed)。
  • --timeout SECONDS:HTTP 超时秒数;0 表示一直等待(默认:0)。
  • --encoding STRING:输入文件编码(默认:utf-8)。

输出约定

输出分两部分:

JSON 结构

{
  "overall_grade": "需关注",
  "summary": "本次体检发现代谢相关多项异常,包括血压偏高、血脂升高、血糖偏高及轻度脂肪肝,需积极改善生活方式并尽快就医评估。",
  "key_findings": [
    {
      "category": "心血管/代谢",
      "finding": "血压142/90mmHg,达高血压1级标准;TG、TC均升高;空腹血糖6.2mmol/L(糖耐量减低范围)",
      "severity": "中度异常",
      "action_needed": "就医复诊,评估是否需要药物干预"
    },
    {
      "category": "消化/代谢",
      "finding": "ALT轻度升高,腹部超声提示轻度脂肪肝",
      "severity": "轻度异常",
      "action_needed": "定期监测,控制体重和饮食"
    }
  ],
  "normal_systems": ["血常规正常", "尿常规正常", "胸部CT未见异常"],
  "priority_actions": [
    {"rank": 1, "action": "前往内科/心内科就诊,评估高血压及代谢综合征", "reason": "血压达高血压标准,合并血脂、血糖异常"},
    {"rank": 2, "action": "3个月内复查肝功能、血脂、血糖", "reason": "监测代谢指标变化"}
  ],
  "lifestyle_advice": ["减少高脂高糖饮食", "每周至少150分钟中等强度有氧运动", "减重至正常BMI范围"],
  "next_exam_suggestion": "6个月后复查,重点关注:血压、血糖、血脂、肝功能"
}

整体解读:以"【整体解读】"开头,用通俗语言面向受检者说明健康状况和建议。

依赖

运行环境

  • Python 3.7+(仅使用标准库,无需额外安装)

外部 API

  • 内部医疗大模型:https://maas-api.hivoice.cn/v1/chat/completions

备注

  • overall_grade 评定标准:全部正常→优秀;有轻微异常但无需就医→良好;有需就诊异常→需关注;有紧急处理项→需重视
  • 发布约束:示例输入、运行输出均放在 skill 包外(data/runs/),skill 目录内仅保留可发布的核心文件
Usage Guidance
Treat this as an incomplete review: the local inspection commands failed with a sandbox error, so the skill should be reviewed again from readable artifacts before installation.
Capability Analysis
Type: OpenClaw Skill Name: unisound-overall-report Version: 1.0.0 The skill is a legitimate tool designed to interpret medical examination reports using a specialized LLM API (maas-api.hivoice.cn). The implementation in `scripts/run.py` is clean, using only Python standard libraries to handle file I/O and HTTP requests, with no evidence of data exfiltration, obfuscation, or unauthorized command execution. The instructions in `SKILL.md` are consistent with the code's behavior and include appropriate privacy and medical boundary disclaimers.
Capability Assessment
Purpose & Capability
Artifact coherence could not be verified because command execution failed before metadata.json or artifact files could be inspected.
Instruction Scope
Runtime instructions could not be reviewed from the supplied files due the local sandbox command failure.
Install Mechanism
Install specifications could not be inspected, so no evidence-backed install risk finding is reported.
Credentials
Requested environment access could not be compared against the skill purpose without readable artifacts.
Persistence & Privilege
No artifact evidence of persistence or privilege behavior was available from the failed inspection.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install unisound-overall-report
  3. After installation, invoke the skill by name or use /unisound-overall-report
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release of health-exam-overall-report: - Provides an overall interpretation of physical exam reports, including overall health grade, key findings with severity, priority actions, and personalized lifestyle recommendations. - Supports both plain text and JSON report input formats. - Outputs JSON and a plain-language interpretation for recipients. - Includes strict data privacy, security practices, and clear medical boundaries. - Offers a command-line interface with configurable parameters for input, output, and model access. - No extra Python dependencies required; uses standard library only.
Metadata
Slug unisound-overall-report
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is unisound-overall-report?

体检报告整体解读。输入完整体检报告文本,输出整体健康评级、主要发现(含严重程度)、优先行动清单和个性化生活方式建议(JSON + 通俗易懂的整体解读)。 It is an AI Agent Skill for Claude Code / OpenClaw, with 108 downloads so far.

How do I install unisound-overall-report?

Run "/install unisound-overall-report" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is unisound-overall-report free?

Yes, unisound-overall-report is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does unisound-overall-report support?

unisound-overall-report is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created unisound-overall-report?

It is built and maintained by Unisound-LLM (@unisound-llm); the current version is v1.0.0.

💬 Comments