ClawHub
Back to skill

Security audit

AI-Cardiac-Rehab

Security checks across malware telemetry and agentic risk

Overview

This is a local cardiac rehabilitation web app, with a privacy documentation gap because its pages load styling and chart assets from a third-party CDN.

Install only on a trusted device, set a strong SECRET_KEY before running, and treat the medical recommendations as support for clinician-reviewed care. If privacy or compliance requirements are strict, self-host Bootstrap and Chart.js or block external CDN access before entering real patient data.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • 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 (2)

Description-Behavior Mismatch

Medium
Confidence
97% confidence
Finding
The application claims patient data is stored locally and not uploaded to the cloud, but the rendered HTML loads Bootstrap and Chart.js from remote CDNs. Even if form contents are not explicitly sent to those providers, every page load discloses client IP address, timing, User-Agent, and referrer metadata to third parties, which undermines the privacy claim in a medical context.

Intent-Code Divergence

Medium
Confidence
98% confidence
Finding
This is a real privacy/security issue because the skill documentation explicitly assures users that data is not uploaded to the cloud, yet browsers are instructed to contact external CDN domains. In a cardiac rehabilitation app, even indirect third-party requests can leak sensitive usage context and create supply-chain risk if the CDN asset is modified or unavailable.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.