Xpersona
U

unisound-diagnosis-review

CLAWHUB · Unverified

诊断编码审核。输入结构化病案 record 与待审核诊断列表，输出诊断编码规则命中、病历证据和审核结论。 Skill: unisound-diagnosis-review Owner: unisound-llm Summary: 诊断编码审核。输入结构化病案 record 与待审核诊断列表，输出诊断编码规则命中、病历证据和审核结论。 Tags: latest:1.0.2 Version history: v1.0.2 | 2026-07-16T02:35:20.808Z | user - Removed the file skill-card.md. - Updated the default model name from u1-insuremed to u2-med in documentation and CLI parameter descriptions. - No functional changes to code or CLI behavior; change is documentation-only. - Ski

62
Rank
84
Safety
976
Downloads

About

What it does, and when to use it.

Capability contract not published. No trust telemetry is available yet. 976 downloads reported by the source. Last updated 10/11/2026.

Avoid when

  • Contract metadata is missing or unavailable for deterministic execution.

Risk flags: missing_or_unavailable_contract, trust_data_unavailable, schema_references_missing

Install and run

Setup complexity: low.

clawhub skill install s1770nyybxj4gn0vk38n6wr6k584x8ea:unisound-diagnosis-review
  1. Setup complexity is LOW. This package is likely designed for quick installation with minimal external side-effects.
  2. Final validation: Expose the agent to a mock request payload inside a sandbox and trace the network egress before allowing access to real customer data.

Contract: missing

curl -s "https://www.xpersona.co/api/v1/agents/clawhub-unisound-llm-unisound-diagnosis-review/snapshot"

Run-check

$0.02 USD

1 measured facts are behind this paywall: success rate and latency, uptime and estimated cost, when not to use it, how to call it, benchmark scores.

Agents pay $0.02 in USDC. A card payment is $0.50, the smallest a card allows.

Documentation

CLAWHUB

24,475 characters of source documentation, loaded on request.

Extracted files

3 files captured from the source.

SKILL.md

---
name: med-diagnosis-review
description: 诊断编码审核。输入结构化病案 record 与待审核诊断列表，输出诊断编码规则命中、病历证据和审核结论。
metadata:
  {
    "openclaw":
      {
        "emoji": "🩺"
      }
  }
---

# 诊断编码审核

概述
----
给定结构化病案 `record` 和待审核诊断列表，本技能从诊断编码审核规则库读取规则，核对病历文书证据，输出诊断编码审核结果。

本技能会：
- 从 `diagnosis_audit_guidelines` 读取诊断编码审核规则。
- 优先按用户提供的 `diagnoses` 候选列表审核；未提供时，默认审核 `record.diagnosis.primaryDiagnosis` 和 `record.diagnosis.otherDiagnoses` 中的全部诊断。
- 调用 skill 内置的 `review_diagnosis_record` function 完成审核。
- 通过 `scripts/run.py` 提供统一 CLI 入口；不转发、不调用当前项目已有 API 服务，也不导入当前项目 `app.*` 模块。

数据安全、隐私与伦理声明
------------------------
- **最小必要原则**：仅处理审核所需的病案文书、诊断编码和诊断名称。
- **严格脱敏**：调用方应在传入前完成姓名、证件号、手机号、详细地址等可识别身份信息的脱敏。
- **不做本地持久化**：本技能不把请求体、中间结果或审核结果写入本地文件或数据库。
- **规则库服务用途**：规则库 API 仅用于读取诊断编码审核规则。
- **模型调用说明**：默认使用内部医疗大模型生成审核说明；鉴权 `appkey` 必须由调用方传入。如需完全离线规则回退，可传 `use_llm=false`。
- **医疗边界**：输出为医保编码审核辅助信息，不构成医疗诊断或治疗建议。

输入格式
--------
统一入口支持 `pdf/doc/docx/xls/xlsx/csv/txt/json`。JSON 可为结构化病案；其他格式会先预处理为一份完整病历文书。未显式传入 `diagnoses` 时，会自动审核结构化病案中的主诊断和全部其他诊断；普通病历文件需要通过参数传入待审核诊断列表。

```json
{
  "record": {
    "hospitalId": "988",
    "serialNum": "0001092574",
    "docs": [
      {
        "docName": "出院记录",
        "fileName": "出院记录",
        "docClassName": "出院记录",
        "content": "..."
      }
    ],
    "diagnosis": {
      "primaryDiagnosis": {"name": "2型糖尿病", "code": "E11.901"},
      "otherDiagnoses": []
    }
  },
  "diagnoses": [
    {"role": "primary", "code": "E11.901", "name": "2型糖尿病"}
  ],
  "appkey": "由平台分配的鉴权 key",
  "model": "",
  "use_llm": true
}
```

字段说明：
- `record`：**必填**。结构化病案 JSON。
- `diagnoses`：可选。待审核诊断候选数组；每项包含 `role`、`code`、`name`。
- `role`：`primary` 表示主诊断；其他值按其他诊断处理。
- `appkey`：**必填**。内部医疗大模型鉴权 key，由平台分配；调用时使用 Bearer 鉴权。
- `model`：可选。内部医疗大模型名称，默认 `u2-med`。
- `use_llm`：可选，默认 `true`。传 `false` 时不调用模型，仅使用本地回退审核逻辑。

审核规则
--------
- **候选优先**：只审核输入候选或病案首页中已有诊断，不自由生成候选列表外的诊断名称或编码。
- **医学内涵优先**：匹配依据以医学内涵、医保目录归类、编码规则和诊疗目的为准，不能只看字面相似。
- **证据优先**：审核结论必须基于病历文书明确支持的信息；不能把病历中未明确支持的信息当作事实。
- **角色边界**：主诊断优先核对本次住院主要诊疗原因；其他诊断核对是否有明确并存、影响诊疗或资源消耗的依据。
- **父类/子类边界**：编码可按规则库从细到粗检索，但父类规则不能自动证明子类编码成立；细分部位、病因、并发症、分型不明确时应输出 `待人工复核`。
- **同义词/简称边界**：仅当简称、同义词与候选诊断在医学内涵和编码归类上明确一致时可视为匹配。
- **模糊保守**：字面相似但医学内涵不同、疑似/排除诊断、仅有既往史或家族史、缺少直接证据时，不强行通过。
- **DRG/CC 边界**：涉及 DRG 分组或严重合并症/并发症、一般合并症/并发症、无合并症/并发症判断时，必须有病历明确支持；不能用模型常识自动升级。

快速开始
--------

```bash
# JSON 结构化病历；未显式传入候选时审核主诊断和全部其他诊断
python3 doctor/icd-drg/diagnosis-review/scripts/run.py \
  --input doctor/icd-drg/diagnosis-review/example/10109_A5204171_1.json \
  --appkey <your-appkey> \
  --no-llm

# TXT/PDF 等普通病历文件；必须传入待审核诊断，可重复传入主诊断和其他诊断
python3 doctor/icd-drg/diagnosis-review/scripts/run.py \
  --input /path/to/record.txt \
  --appkey <your-appkey> \
  --diagnosis 'primary|E11.901|2型糖尿病' \
  --diagnosis 'other|I10.x00|高血压' \
  --save-prepared
```

参数说明
--------
- `--input PATH`：**必填**。结构化病案 JSON，或包含 `record` 的请求体 JSON。
- `--input-type auto|pdf|doc|docx|xls|xlsx|csv|txt|json`：输入类型，默认 `auto`。
- `--sheet STRING`：读取 Excel 时指定 sheet（可选）。
- `--encoding STR

_meta.json

{
  "ownerId": "kn7ft98qwhxyrgcy7b7rzdhrrd8366k6",
  "slug": "unisound-diagnosis-review",
  "version": "1.0.2",
  "publishedAt": 1784169320808
}

skill-card.md

## Description:

Reviews structured medical records and candidate diagnoses to report diagnosis-coding rule matches, supporting record evidence, and an audit conclusion.

This skill is ready for commercial/non-commercial use.

## Publisher:

[unisound-llm](https://clawhub.ai/user/unisound-llm)

### License/Terms of Use:

MIT-0

## Use Case:

Healthcare coding reviewers and developers use this skill to audit supplied diagnosis codes against medical-record evidence and diagnosis audit rules. Its output is coding-audit assistance and not medical diagnosis or treatment advice.

### Deployment Geography for Use:

Global

## Known Risks and Mitigations:

Risk: Sensitive medical-record content and bearer credentials can be sent to an LLM endpoint.

Mitigation: Use only approved healthcare endpoints and keep LLM use disabled unless that endpoint and data flow are authorized.

Risk: Untrusted callers can select the LLM base URL.

Mitigation: Do not allow untrusted callers to set --base or payload.base; restrict execution to approved endpoint configuration.

Risk: Prepared record text can be persisted when requested.

Mitigation: Avoid --save-prepared for PHI and pass only authorized, de-identified patient data.

## Reference(s):

- [ClawHub skill page](https://clawhub.ai/unisound-llm/skills/unisound-diagnosis-review)

## Skill Output:

**Output Type(s):** [text, JSON, guidance]

**Output Format:** [JSON object with final_decision and reasoning fields]

**Output Parameters:** [1D]

**Other Properties Related to Output:** [final_decision is one of: 通过, 不通过, 待人工复核.]

## Skill Version(s):

1.0.2 (source: server release evidence)

## Ethical Considerations:

Users should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.

Sponsored

Ads related to unisound-diagnosis-review and adjacent AI workflows.