ClawHub
Back to skill

Security audit

unisound-pe-missing-negative

Security checks across malware telemetry and agentic risk

Overview

This skill coherently performs user-directed medical-record quality checks, but users should treat its remote API call and saved outputs as sensitive.

Install only if you are allowed to send de-identified medical record text to the configured HiVoice MaaS endpoint. Avoid using --save-prepared with sensitive records unless you need the debug file, and delete output files according to your retention policy.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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
Findings (2)

Lp3

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding
The skill documentation exposes capabilities to read files, write local output, and send record content over the network to an external MaaS endpoint, but it does not declare corresponding permissions. This creates a trust and review gap: operators may approve or run the skill without understanding that sensitive medical text can be exfiltrated or persisted locally.

Intent-Code Divergence

Medium
Confidence
97% confidence
Finding
The file claims there is no local persistence and that data is destroyed after the call, yet the same skill defines a default output file and an option to save prepared text. For medical records, this mismatch is risky because users may rely on the privacy statement while the tool actually leaves PHI-sensitive artifacts on disk.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.