zepp-os-sensors
Pass
Audited by Gen Agent Trust Hub on Aug 10, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill serves as a technical reference for Zepp OS sensor integration. It correctly identifies the necessary permissions (e.g.,
data:user.hd.heart_rate,device:os.accelerometer) and follows the platform's standard 'instantiate → subscribe → read → stop' pattern. - [DATA_EXPOSURE]: While the skill facilitates access to sensitive data such as heart rate, blood oxygen levels, and geolocation, it does so through the official
@zos/sensorAPI. The documentation emphasizes declaring these permissions in the application configuration (app.json), which is a standard security practice for the platform. There are no instructions or code snippets that attempt to exfiltrate this data to unauthorized external servers. - [REMOTE_CODE_EXECUTION]: No evidence of external script downloads, package installations from untrusted sources, or dynamic code execution was found. All examples use local imports from the expected platform modules.
- [INDIRECT_PROMPT_INJECTION]: The skill defines patterns for ingesting sensor data. While sensor data constitutes an external input, the risk of indirect prompt injection in this context (embedded watch device) is negligible and handled by the underlying OS/LLM guardrails. No capability for executing instructions hidden within sensor data was identified.
Audit Metadata