← Back to Skills Marketplace
unisound-llm

unisound-chronic-injury

by Unisound-LLM · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ Security Clean
107
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install unisound-chronic-injury
Description
理赔陈旧伤审核。由调用方传入完整题干(常见为影像报告 + 指定部位 + 输出格式),经内部医疗大模型判断伤情类别;仅含 scripts/run.py,无 _shared 依赖。
README (SKILL.md)

陈旧伤审核

概述

面向理赔材料中的影像/检查报告文字(可由 OCR 或人工整理),在调用方给定的分类体系下(常见为 陈旧伤 / 新伤 / 无伤 / 未提及),对指定解剖部位做出伤情类别判断,并按题干要求的固定行格式输出。

本 skill 不接院内 PACS,题干由上游拼装;与慢病/大病审核(走后端业务接口)不同,本包直连公司内部医疗大模型,接口与参数约定同 medbench 类评测脚本。

数据与合规

  • 送入模型前应对姓名、证件号、影像号等做脱敏(由调用方完成);本脚本不落盘用户内容(除非使用 --output 显式保存结果)。
  • 输出为辅助审核意见,不构成法医鉴定或最终理赔决定。

题目输入(三选一)

  1. --question:完整题干字符串。
  2. --input PATH
    • .json / .jsonl:根为对象,或仅含 1 个对象的 JSON 数组;对象须含非空 inputquestion
    • 其他扩展名:整文件 UTF-8 纯文本作为题干。
    • - 时从 stdin 读。
  3. stdin(未传 --question / --input、且非 TTY):规则同 -

可选字段 meta / other:任意 JSON 对象,原样进入运行结果中的 meta

题干与输出格式(由调用方约定)

典型题干中会包含:

  • 影像检查报告摘录(影像描述 / 影像诊断等)
  • 指定检查部位与所见依据
  • 待选类别列表
  • 严格输出模板(例如两行:伤情类别:… / 判断依据:…

模型须完全遵守题干中的格式与类别用词;默认系统提示词与上述约束对齐。

运行结果(stdout)

默认可解析 JSON:skill陈旧伤审核statusokanswer 为模型文本,question 为本次用户消息全文,meta 等为透传/溯源字段(含义同 doctor/icd-drg/drg-grouping 等脚本)。

用法示例

python3 scripts/run.py --input case.json --appkey YOUR_KEY
python3 scripts/run.py --input case.json --text-only --appkey YOUR_KEY

参数

  • --question / --input:与 stdin 三选一(二者不可同时)。
  • --appkey STRING必填--dry-run 除外)。
  • --index--match-id--batch:仅 .jsonl
  • --dry-run--model--api-url--temperature--timeout--system-prompt--output--text-only:同仓库内其他直连大模型脚本。

模型

  • 默认 u1-insuremedhttps://maas-api.hivoice.cn/v1/chat/completions
Usage Guidance
Treat this as an inconclusive low-confidence review and rerun with readable metadata.json and artifact/ contents before installing.
Capability Analysis
Type: OpenClaw Skill Name: unisound-chronic-injury Version: 1.0.0 The skill is a medical injury review tool designed to classify injuries based on imaging reports by calling a remote LLM API (hivoice.cn). The implementation in scripts/run.py uses standard Python libraries to process local input files and perform authorized API requests, with no evidence of malicious intent, unauthorized data exfiltration, or hidden execution logic.
Capability Assessment
Purpose & Capability
metadata.json and artifact/ were not readable because local command execution failed, so purpose and capability coherence could not be verified from artifacts.
Instruction Scope
Instruction scope could not be assessed without reading SKILL.md or related artifact files.
Install Mechanism
Install mechanism could not be assessed because install specs and manifests were unavailable to this review.
Credentials
Environment access proportionality could not be checked without artifact contents.
Persistence & Privilege
Persistence and privilege behavior could not be checked without artifact contents.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install unisound-chronic-injury
  3. After installation, invoke the skill by name or use /unisound-chronic-injury
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
med-chronic-injury-review 1.0.0 - Initial release for chronic injury claim review based on medical text reports. - Supports classification of injury status (e.g., chronic injury, acute injury, no injury, not mentioned) for specified body parts, with strict adherence to provided formats. - Operates via direct integration with internal medical large language model; no PACS connection. - Requires pre-anonymized input from caller; does not store user data unless explicitly output. - Flexible input via command-line argument, file, or stdin; supports JSON formats. - Output includes machine-readable JSON and model-generated classification text.
Metadata
Slug unisound-chronic-injury
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is unisound-chronic-injury?

理赔陈旧伤审核。由调用方传入完整题干(常见为影像报告 + 指定部位 + 输出格式),经内部医疗大模型判断伤情类别;仅含 scripts/run.py,无 _shared 依赖。 It is an AI Agent Skill for Claude Code / OpenClaw, with 107 downloads so far.

How do I install unisound-chronic-injury?

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

Is unisound-chronic-injury free?

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

Which platforms does unisound-chronic-injury support?

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

Who created unisound-chronic-injury?

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

💬 Comments