ClawHub
Back to skill

Security audit

Need To Parse Complex Medical Dicom File

Security checks across malware telemetry and agentic risk

Overview

This skill is a minimal prompt-only DICOM parsing helper with vague scope but no executable code, persistence, credential use, or data export behavior.

Install only if you are comfortable with a very general DICOM-related reasoning aid. Because medical images can contain sensitive health information, use it only on files the user has explicitly provided or authorized, avoid sending DICOM content to external services unless separately approved, and add stronger workflow-specific safeguards if clinical or regulated data is involved.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Vague Triggers

Medium
Confidence
89% confidence
Finding
The activation condition is extremely broad and vague: it triggers whenever the agent perceives a need to parse complex medical DICOM files, without defining scope, prerequisites, trust boundaries, or required user authorization. In an autonomous agent setting, this can cause unintended invocation on sensitive medical data or in workflows the skill was not designed to handle, increasing the risk of privacy violations, unsafe processing, or incorrect downstream actions.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.