ClawHub
Back to skill

Security audit

unisound-medical-term-normalization

Security checks across malware telemetry and agentic risk

Overview

The skill performs the stated medical-record normalization task, but users should treat inputs as sensitive because records are sent to a remote model and optional files can persist data locally.

Install only if you are comfortable sending de-identified medical record text to the configured model endpoint with your app key. Remove names, IDs, phone numbers, addresses, and other identifiers before use, and avoid --save-prepared or output-file flags unless you intend to store the resulting medical text on disk.

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 (3)

Lp3

Medium
Category
MCP Least Privilege
Confidence
96% confidence
Finding
The skill documentation describes capabilities to read local files, optionally write outputs/preprocessed text, and send medical record content to a remote API, yet it declares no explicit permissions. This creates a governance gap: operators may invoke the skill without clear visibility or policy enforcement around sensitive file and network access, which is especially risky because the processed content is medical data and the skill can persist derived text when flags are used.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The script sends full medical-record content to a remote LLM endpoint via `call_llm`/`_http_post` without any explicit consent flow, warning, minimization, or redaction step. Because the data is highly sensitive health information, undisclosed transmission to an external service materially increases privacy, compliance, and breach risk even if the endpoint is described as an internal medical model.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
When `--save-prepared` is used, the script writes preprocessed medical text directly to disk, potentially preserving sensitive patient data in plaintext under a predictable runs/debug directory. This creates a local confidentiality risk through leftover files, weak filesystem permissions, backups, or accidental sharing, especially in multi-user or managed environments.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.