Birth Medical Cert Ocr

• Outcome: Review
• Latest audit: Jul 9, 2026 · 10:06 AM UTC
• Version: 1.0.4

Overview

This skill performs remote OCR, but it needs review because it handles highly sensitive birth-certificate data and its metadata/code allow broader OCR use than the narrow description claims.

Install only if you are comfortable sending birth-certificate images and extracted family/medical identity data to Scnet's remote OCR service. Confirm you have legal authority or guardian consent, keep the API key in a protected local env file, avoid using it for general OCR, and prefer a version that hardcodes or validates BIRTH_CERTIFICATE before upload.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (9)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 89% confidence
• Finding: The skill documentation indicates the capability to read local files, invoke shell commands, and send data over the network, but it does not declare permissions accordingly. This creates a transparency and consent problem: a caller may not realize the skill can exfiltrate sensitive local documents such as birth certificates to a remote service. In this context, the risk is heightened because the target data contains highly sensitive personal and medical information about minors and parents.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 96% confidence
• Finding: The skill claims it is strictly limited to birth medical certificate OCR and only triggers when users mention specific keywords, yet the described behavior allows arbitrary command-line ocrType input, lacks content validation, and does not implement the trigger restriction. This mismatch can enable use as a broader OCR/exfiltration wrapper than advertised, undermining user expectations and policy controls around highly sensitive identity documents.

Description-Behavior Mismatch

Medium

• Confidence: 98% confidence
• Finding: The skill description says it must only handle birth medical certificate OCR, but the code accepts any user-supplied ocrType from argv and forwards it directly to the backend API without validation. This creates a scope-bypass that can turn a narrowly permitted document-recognition skill into a generic OCR proxy, increasing the chance of unauthorized processing of other document types and sensitive data.

Description-Behavior Mismatch

High

• Confidence: 98% confidence
• Finding: The manifest description explicitly broadens the skill from a narrowly scoped birth medical certificate OCR tool to general image/text OCR ('识别图片中的文字'), which conflicts with the stated restriction that it must only trigger for birth-certificate-related requests. This kind of scope expansion is dangerous because agents often rely on manifest metadata for routing and authorization decisions, so an overly broad description can cause the skill to be invoked on unrelated sensitive documents and bypass policy boundaries.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: This README documents a skill for OCR of birth medical certificates, which inherently involves highly sensitive personal and medical data, yet it provides no warning about privacy, consent, retention, or restricted handling. In this context, the omission increases the risk of misuse, over-collection, or deployment without adequate safeguards, especially because the skill is explicitly intended to extract names, dates, parent data, and ID numbers.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The documentation describes uploading highly sensitive identity and medical documents containing newborn, parent, address, and ID-number data to a third-party OCR endpoint, but provides no privacy, retention, consent, residency, or handling guidance. In the context of a skill specifically designed to process birth medical certificates, this omission is dangerous because it normalizes external transmission of regulated personal data without safeguards or user awareness.

External Transmission

Medium

• Category: Data Exfiltration

Content

| 变量名 | 默认值 | 说明 |
|--------|--------|------|
| SCNET_API_KEY | 必需 | Scnet API 密钥 |
| SCNET_API_BASE | https://api.scnet.cn/api/llm/v1 | API 基础地址（一般无需修改） |

### 输出

• Confidence: 93% confidence
• Finding: https://api.scnet.cn/

External Transmission

Medium

• Category: Data Exfiltration

Content

# Sugon-Scnet OCR API 文档摘要

## 接口地址
`POST https://api.scnet.cn/api/llm/v1/ocr/recognize`

## 请求头
- `Content-Type: multipart/form-data`

• Confidence: 89% confidence
• Finding: https://api.scnet.cn/

External Transmission

Medium

• Category: Data Exfiltration

Content

)
        sys.exit(error_msg)

    config.setdefault('SCNET_API_BASE', 'https://api.scnet.cn/api/llm/v1')
    return config

def recognize_with_retry(ocr_type, file_path, config, retry_count=0):

• Confidence: 88% confidence
• Finding: https://api.scnet.cn/

VirusTotal

64/64 vendors flagged this skill as clean.

Static analysis

No suspicious patterns detected.