unisound-drg-grouping

ReviewAudited by ClawScan on May 15, 2026.

Overview

This skill appears to do what it claims—send user-provided DRG case text to a disclosed medical model API—but users should protect patient data and the app key.

Before installing or using this skill, confirm that sending the case text to the configured model API is allowed, de-identify real patient data when required, keep the app key private, and avoid overriding the API URL unless the endpoint is trusted and approved.

Findings (4)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

Case details may leave the local environment and be processed by the configured model provider.

Why it was flagged

The user-provided case prompt is sent to the disclosed external/internal model API. This is expected for the skill, but the prompt may contain medical or patient information.

Skill content
DEFAULT_API_URL = "https://maas-api.hivoice.cn/v1/chat/completions" ... {"role": "user", "content": user_prompt}
Recommendation

Use only with approved endpoints, de-identify real patient data where required, and follow the organization’s medical-data handling rules.

What this means

Anyone who obtains the app key may be able to use the associated model service account or quota.

Why it was flagged

The script requires an app key for non-dry-run use and sends it as a bearer token to the model API. This is purpose-aligned but is still a credential.

Skill content
p.add_argument("--appkey", default="", help="内部医疗大模型鉴权 key。") ... "Authorization": f"Bearer {appkey}"
Recommendation

Keep the app key secret, avoid sharing command history or logs containing it, and rotate it if exposure is suspected.

What this means

If the API URL is changed to an untrusted service, the app key and medical prompt text could be sent there.

Why it was flagged

The API endpoint is user-configurable, and the same request includes both the prompt data and bearer credential. This is disclosed, but an untrusted endpoint would receive sensitive data.

Skill content
p.add_argument("--api-url", default=DEFAULT_API_URL, help="OpenAI 兼容接口地址") ... Request(api_url, ... "Authorization": f"Bearer {appkey}")
Recommendation

Do not let untrusted task text or third parties choose the API URL; use the default or another approved endpoint only.

What this means

Users have less assurance that the package identity and registry listing are perfectly aligned.

Why it was flagged

The internal metadata differs from the submitted registry identity shown for this review. This does not indicate malicious behavior, but it slightly weakens provenance clarity.

Skill content
"ownerId": "kn76wejkeqxfc03j0rfxp2jaj982m7aa", "slug": "doctor.icd-drg.drg-grouping"
Recommendation

Verify the publisher and package identity before relying on it in a clinical or production workflow.