ClawHub
Back to skill

Security audit

unisound-med-record-qc

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed medical-record QC helper that sends user-provided text to a stated medical model API and writes requested outputs, so the main risk is handling sensitive patient data carefully.

Install only in an environment approved for the medical data you plan to process. De-identify real patient records before use, confirm the configured API endpoint and appkey handling meet your compliance requirements, and be careful with stdout logs and --output files because they can contain the original question, metadata, and model answer.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Lp3

Medium
Category
MCP Least Privilege
Confidence
83% confidence
Finding
The skill documentation describes a script that can read local files, write output files, and make network requests to an external medical-model API, yet it declares no permissions. This creates a transparency and governance gap: deployers may underestimate the skill's access to sensitive medical data and allow it to process protected records without explicit review or sandboxing.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The script sends user-supplied medical record or report content directly to a remote API endpoint, and the CLI help text does not clearly warn operators that potentially sensitive clinical data will leave the local environment. In a medical QC skill, prompts are likely to contain PHI/PII and regulated health data, so silent transmission to an external service creates a real confidentiality and compliance risk even if the endpoint is intended.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The script persists the full question, metadata, and model answer to stdout and optionally to output files, with no warning that these fields may contain sensitive medical information. In this skill's context, records and metadata may include patient identifiers or case details, so writing them verbatim increases the chance of accidental retention, logging, sharing, or downstream exposure.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.