ClawHub
Back to skill

Security audit

unisound-similar-case-retrieval

Security checks across malware telemetry and agentic risk

Overview

The skill is a disclosed clinical research helper that sends user-provided, de-identified case summaries to a named medical model endpoint for ranking, with privacy caution needed before use.

Install only if your workflow is allowed to send de-identified clinical summaries to the documented remote model endpoint. Do not provide identifiable patient data unless your organization has approved that processor, retention policy, app key handling, and ethics/compliance requirements.

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 (2)

Lp3

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding
The skill declares command-line execution, reads input files, optionally writes output files, and makes outbound network calls to an external model endpoint, yet no explicit permissions are declared. This creates a governance and sandboxing gap: operators cannot reliably understand or restrict what the skill can access, and in a medical-data context that increases the risk of unintended PHI exposure or overly broad runtime privileges.

Missing User Warnings

High
Confidence
98% confidence
Finding
The script sends anchor_case and candidate_cases, which are clinical case summaries and may contain sensitive health information, to an external remote API endpoint. There is no explicit consent notice, data-classification check, de-identification step, or transport policy guard in the workflow, creating a real privacy and compliance risk in a medical context.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.