Back to skill

Security audit

Apple Health Data AI Analyzer

Security checks for vulnerabilities and agentic risk

Overview

The scanners rate this health analysis skill benign with high confidence, reporting no major security issues and one low probability finding judged expected for its purpose.

The scan completed with full coverage of the two reported files and supports the benign verdict. Before installing, consider whether you are comfortable with a skill that reads health information from a local API; the scanner finding highlights a concern about not asking the user, though it was judged expected here. Benign does not mean risk free.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
80% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 29)May include surrounding context.

md
- Treat the local Health Data AI Analyzer API as the source of truth.
- Read only the minimum localhost endpoint needed to answer the request.
- Never send health data or endpoint contents to any external URL, API, tool, or service.
- Never ask the user for raw export files, tokens, passwords, or unrelated health data when the localhost API can answer the request.
- Do not invent missing data.
- Do not give medical advice or diagnoses.
- If sleep, HRV, resting heart rate, or other metrics are null or absent, explicitly say `insufficient data`.