Security audit
Security checks for vulnerabilities and agentic risk
The skill is mostly transparent, but it guides automated sharing of live Garmin location and sensor data without enough privacy, consent, or scoping safeguards.
Only install this if you are prepared to add your own guardrails: get explicit consent from the tracked user, send the minimum necessary fields, use trusted authenticated endpoints, avoid unofficial session-token scraping for anything important, and define retention and access controls before forwarding live location or sensor data.
The declared description promises a functional Garmin real-time tracking and data delivery capability, but the code chunk does not implement any of that behavior. It is a stub/example script with a print statement and TODO comments only. This is a material mismatch in primary purpose and implemented capabilities.
The skill explicitly frames live location, activity, and sensor streaming to external systems without any accompanying privacy, consent, retention, or third-party sharing safeguards. Because location and biometric/activity streams are highly sensitive, presenting transmission guidance without clear warnings or consent requirements increases the risk of unauthorized surveillance, privacy violations, or mishandling of regulated health-adjacent data.
The activation text is broad enough that an agent could invoke this skill for common Garmin-related requests involving location, health, or sensor data without sufficient narrowing conditions. In a data-sensitive context, over-broad routing can cause the system to surface instructions for collecting or transmitting sensitive telemetry when the user did not explicitly request that level of access or sharing.
No suspicious patterns detected.
