← Back to Skills Marketplace
257
Downloads
0
Stars
1
Active Installs
2
Versions
Install in OpenClaw
/install critical-disease-review
Description
重大疾病理赔智能评估（支持 28 种病种）。输入住院病历结构化数据，调用内网评估接口，输出原始 JSON 与自然语言结论（结论 + 证据）。
README (SKILL.md)
重大疾病理赔评估
概述
给定一份住院病历（结构化 medicalRecord，含诊断/文书等），本技能调用内网评估服务，对 任意重大疾病病种 进行判定，并返回：
- 最终结论：符合/不符合（来自
final_result） - 原因与证据：按条件逐条给出 evidence（可带来源 source）
- 原始结果：接口返回的完整 JSON（便于追溯与二次加工）
数据安全、隐私与伦理声明
- 最小必要原则：仅处理完成评估所必需的数据字段；不要求提供与评估无关的身份信息。
- 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（如姓名、证件号、手机号、详细地址、人脸/影像等）。仅传递脱敏后的必要信息用于本次 skill 调用。
- 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（包含磁盘文件、数据库、日志）。仅在内存中短暂处理；本次调用结束即销毁。
- 第三方 API 风险提示：在功能需要时，可能会调用第三方模型/服务接口；此时仅会发送脱敏后的必要信息，并使用加密传输。除完成本次请求外，不用于任何其他用途（如训练、画像、营销）。
- 医疗边界：本技能输出为理赔条款/条件匹配与证据整理的辅助信息，不构成医疗诊断或治疗建议；如涉及临床判断请以执业医生意见为准。
支持病种（28）
- 心脏瓣膜手术：
heart_valve_surgery - 主动脉手术：
aortic_surgery - 冠状动脉搭桥术：
coronary_artery_bypass - 重大器官或造血干细胞移植术：
major_organ_transplant - 恶性肿瘤——重度：
malignant_tumor - 严重慢性肾衰竭：
severe_chronic_kidney_failure - 严重慢性肝衰竭：
severe_chronic_liver_failure - 急性重症肝炎或亚急性重症肝炎：
acute_severe_hepatitis - 严重慢性呼吸衰竭：
severe_chronic_respiratory_failure - 严重特发性肺动脉高压：
severe_idiopathic_pulmonary_hypertension - 严重脑损伤：
severe_brain_injury - 深度昏迷：
deep_coma - 严重脑中风后遗症：
severe_stroke_sequelae - 严重阿尔茨海默病：
severe_alzheimers_disease - 严重原发性帕金森病：
severe_primary_parkinsons_disease - 严重运动神经元病：
severe_motor_neuron_disease - 严重脑炎后遗症或严重脑膜炎后遗症：
severe_brain_encephalitis_sequelae - 严重非恶性颅内肿瘤：
severe_non_malignant_intracranial_tumor - 瘫痪：
paralysis - 双耳失聪：
bilateral_deafness - 双目失明：
bilateral_blindness - 语言能力丧失：
language_ability_loss - 严重克罗恩病：
severe_crohn_disease - 严重溃疡性结肠炎：
severe_ulcerative_colitis - 重型再生障碍性贫血：
severe_aplastic_anemia - 较重急性心肌梗死：
moderate_acute_myocardial_infarction - 严重Ⅲ度烧伤：
severe_third_degree_burn - 多个肢体缺失：
multiple_limb_loss
输入格式
请求为 JSON（示例见 ../data/med-major-disease-assess/req_data.json），最少需要包含：
medicalRecord：病历结构化对象mainDiagName/otherDiagName：诊断信息（字符串/JSON 字符串均可）docs：文书列表（每个 doc 至少包含docType与文本字段，如format_page_text）
最小校验（先校验，再审核）
脚本会先对入参做最小结构校验，校验通过后才会调用审核接口：
- 请求体必须是 JSON object
- 必须包含
medicalRecord（object） medicalRecord.docs必须是非空数组docs中至少一项包含docType
后端接口
- HTTP API：
http://10.10.20.15:9010/api/v1/assessment/assess/{disease}?model_type=qwq - Content-Type：
application/json
快速开始
# 在本目录下运行
python3 scripts/major_disease_assess.py \
--disease aortic_surgery \
--input ../data/med-major-disease-assess/req_data.json
参数说明
--disease STRING- 病种类型（如：
aortic_surgery、heart_valve_surgery等）。
- 病种类型（如：
--input PATH- 输入请求 JSON 路径（UTF-8）。
--output-json PATH- 保存接口原始返回 JSON（默认：
../runs/med-major-disease-assess/{disease}_resp.json）。
- 保存接口原始返回 JSON（默认：
--output-text PATH- 保存自然语言结论文本（默认：
../runs/med-major-disease-assess/{disease}_resp.txt）。
- 保存自然语言结论文本（默认：
--model-type STRING- 查询参数
model_type（默认：qwq）。
- 查询参数
--timeout SECONDS- HTTP 超时秒数（默认：60）。
输出约定
- 若指定输出路径的父目录不存在，会自动创建。
- 自然语言输出默认包含：
- 结论（符合/不符合）
- 原因（
final_result.reason） - 逐条条件的证据（
conditions[*].evidence，并附source/description）
备注
- 发布约束：示例输入、运行输出、自测脚本均放在 skill 包外（分别位于
../data/、../runs/、../self_tests/），skill 目录内仅保留可发布的核心文件（scripts/、SKILL.md、_meta.json）。
Usage Guidance
Do NOT install or run this skill on real patient data until these discrepancies are resolved. Key issues to clarify or fix before use: (1) Confirm the correct backend endpoint — the README claims an internal IP but the code posts to an external domain; verify ownership and trustworthiness of that host. (2) Implement and show PII redaction logic in code (the README promises desensitization but the script sends the payload unchanged). (3) If no local persistence is truly required, remove or make disk writes optional and document where files are stored; otherwise warn users and require an explicit opt-in. (4) Prefer making the endpoint configurable (not hard-coded) and require an allow-list for allowed hosts. (5) Have the code and documentation audited by a security/privacy reviewer and test in an isolated network with synthetic data. If you cannot obtain clear answers and a trusted internal endpoint, treat this skill as high risk for data exfiltration.
Capability Analysis
Type: OpenClaw Skill
Name: critical-disease-review
Version: 1.0.1
The skill sends sensitive medical record data to a remote API (shangbao.yunzhisheng.cn) for insurance claim assessment. While this is its stated purpose, it is classified as suspicious because the SKILL.md documentation explicitly claims 'strict de-identification' is performed before transmission, yet the provided code in scripts/major_disease_assess.py lacks any de-identification logic and simply forwards the input data. Additionally, there is a discrepancy between the internal IP address (10.10.20.15) mentioned in the documentation and the public domain used in the actual script.
Capability Assessment
Purpose & Capability
The SKILL.md states the skill calls an internal API (example: http://10.10.20.15:9010/...) and emphasizes privacy guarantees. The actual code uses BASE_URL = "https://shangbao.yunzhisheng.cn/skills/critical-disease/api/v1/assessment/assess" (an external domain). This is a direct mismatch: either the docs are outdated/misleading or the code is exfiltrating data to a host not described in the description.
Instruction Scope
SKILL.md promises '严格脱敏' (strict desensitization) and '不做本地持久化' (no local persistence). The code performs only structural validation and then POSTs the payload as-is — there is no PII redaction logic. Additionally, the script writes the raw response JSON and the natural-language output to disk under ../runs/..., contradicting the 'no local persistence' claim.
Install Mechanism
No install spec; this is an instruction + small Python script bundle. No external installer downloads or package installs are performed by the skill itself. Risk here is limited to what the Python code does at runtime (network I/O and file writes).
Credentials
The skill declares no required credentials or env vars, yet it sends the provided medicalRecord payload to a remote HTTPS endpoint. Because no credentials are required, the code will happily transmit potentially sensitive medical data off-host to the hard-coded external domain — this is disproportionate to the advertised 'internal assessment' purpose and the stated privacy guarantees.
Persistence & Privilege
The SKILL.md asserts 'no local persistence', but the script creates directories and writes raw response JSON and text files to ../runs/med-major-disease-assess by default. While 'always' is false and the skill is not forced, the local file writes are a contradiction and increase the risk surface (sensitive outputs persisted to disk).
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install critical-disease-review - After installation, invoke the skill by name or use
/critical-disease-review - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.1
- 首个版本发布，无功能变更
- 支持输入结构化住院病历，自动进行28种重大疾病理赔条款的判定与证据整理。
- 内置数据安全、隐私保护与使用边界说明。
- 支持保存接口原始 JSON 及自然语言评估结论，便于追溯和二次加工。
- 提供命令行快速起步和详细参数说明。
v1.0.0
med-critical-disease-review 1.0.0
- Initial release with support for 28 major disease categories for insurance claim assessment.
- Accepts structured hospital records as input, performs assessment via internal API, and returns both detailed JSON and natural language conclusions (including supporting evidence and sources).
- Implements strict data privacy, sensitivity filtering, and no local persistence of user or intermediate data.
- Provides validation on input format and minimum required fields.
- Includes clear usage instructions, supported parameters, and output conventions for both JSON and text formats.
Metadata
Frequently Asked Questions
What is critical-disease-review?
重大疾病理赔智能评估（支持 28 种病种）。输入住院病历结构化数据，调用内网评估接口，输出原始 JSON 与自然语言结论（结论 + 证据）。 It is an AI Agent Skill for Claude Code / OpenClaw, with 257 downloads so far.
How do I install critical-disease-review?
Run "/install critical-disease-review" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is critical-disease-review free?
Yes, critical-disease-review is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does critical-disease-review support?
critical-disease-review is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created critical-disease-review?
It is built and maintained by aaiccee (@aaiccee); the current version is v1.0.1.
More Skills