Security audit

Hypertension Manager

Security checks for vulnerabilities and agentic risk

Overview

This is an instruction-only blood pressure logging skill with health-data and medical-adjacent caveats, but no executable code, hidden access, credential use, or exfiltration behavior.

Install only if you are comfortable entering blood pressure readings and basic health details into local daily notes. Treat outputs as informational, not diagnosis or treatment advice, and seek professional care for abnormal readings, symptoms, or medication questions. The audio link opens a third-party site, so do not assume the skill's local-data privacy statement applies after clicking it.

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)

Intent-Code Divergence

Medium

Confidence: 94% confidence
Finding: The skill states it does not recommend or track external services, yet it instructs the agent to provide clickable external audio links tied to a user's health condition and sex. This creates a privacy and trust risk because users may be steered to third-party resources based on sensitive health data, contradicting the stated privacy model and potentially exposing user context to external sites.

Vague Triggers

Medium

Confidence: 77% confidence
Finding: The trigger set is broad enough that the skill may activate during general conversation about blood pressure, causing unsolicited collection or processing of sensitive health information. In a health-related skill, overbroad activation increases privacy and safety risk because users may receive medical-style guidance or data recording flows without clearly opting into this specific skill.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: This reference gives condition-specific blood pressure targets and suggests that persistent readings after medication may indicate the need to adjust treatment, but it does not clearly warn users not to self-diagnose or change medication without a clinician. In a hypertension-management skill, users may treat this as actionable medical advice, increasing the risk of unsafe self-management or delayed professional care.

Static analysis

No suspicious patterns detected.