ClawHub
Back to skill

Security audit

unisound-med-clinical-diagnosis

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed clinical prompt runner that sends user-provided cases to a configured medical-model API, with no hidden persistence or unrelated behavior found.

Install only if you are comfortable sending case text to the configured medical-model API. Do not include patient identifiers or protected health information unless your organization has approved that endpoint and data flow; use --dry-run to inspect parsed input without making a network call.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Lp3

Medium
Category
MCP Least Privilege
Confidence
91% confidence
Finding
The skill documentation describes capabilities to read local files/stdin, write output files, and send prompts to a remote medical-model API, but it does not declare corresponding permissions. This creates a transparency and governance gap: integrators may deploy the skill without understanding that user-supplied content, potentially including sensitive clinical data, can be read from disk and transmitted over the network.

Description-Behavior Mismatch

Medium
Confidence
90% confidence
Finding
The skill is presented as a standalone clinical diagnosis utility, but it forwards potentially sensitive clinical question content to a remote API endpoint. In a medical context, case descriptions can contain protected health information, so undisclosed external transmission creates a real confidentiality and compliance risk even if the destination service is expected.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The code sends user-supplied clinical questions directly to a remote chat completion API without any user-facing warning at the point of use. Because these prompts may include patient data, this can cause unintended disclosure of sensitive medical information and create privacy, contractual, or regulatory exposure.

Static analysis

No suspicious patterns detected.