ClawHub
Back to skill

Security audit

个人健康管家免费版

Security checks across malware telemetry and agentic risk

Overview

The skill is mostly a coherent personal health assistant, but it handles sensitive health data while making local-only privacy claims that are not fully aligned with its required agent LLM use.

Review before installing. Use it only with health data you are comfortable sending to the agent/model provider unless your agent is configured for local processing, and keep its file access limited to the intended ~/.health directory where possible.

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (3)

Description-Behavior Mismatch

Medium
Confidence
90% confidence
Finding
The skill repeatedly claims that all data stays local and the free edition is focused on local-only personal health management, yet the dependency and troubleshooting sections introduce an LLM API requirement and network diagnostics. For a health-data skill, this mismatch can mislead users into exposing sensitive medical information to remote model providers when they reasonably expect strictly local processing.

Intent-Code Divergence

Medium
Confidence
95% confidence
Finding
The documentation states that all data remains local and no external API key is needed, but elsewhere says an LLM API is required. In the context of personal health and medical-report processing, this is a material privacy and trust issue because users may enter sensitive health data under false assumptions that nothing leaves their device.

Vague Triggers

Medium
Confidence
78% confidence
Finding
The trigger condition is overly broad, effectively matching generic requests for analysis, reporting, insights, and visualization. In an agent context with read/write/exec tools enabled, vague activation criteria increase the chance the skill is invoked on unrelated or more sensitive user data than intended, including health-adjacent files or broader personal datasets.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.