camerax
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides patterns for processing untrusted external data from camera sensors, which introduces a potential attack surface for indirect prompt injection.
- Ingestion points: Camera image frames are accessed and processed via
ImageProxyandImageAnalysisbuffers as described inreferences/camera1-to-camerax.md,references/camera2-to-camerax.md, andreferences/mlkit-spatial.md. - Boundary markers: None identified; image data is typically processed as binary buffers by analyzers without inherent delimiters for embedded instructions.
- Capability inventory: The skill provides logic for writing files to storage using
ImageCapture.takePicture(e.g., inSKILL.md) and transmitting data over Bluetooth or Wi-Fi using theWearable.getDataClientAPI as documented inreferences/wear-os.md. - Sanitization: Standard technical patterns for coordinate mapping and image processing are provided, but there are no specific filters or sanitization steps mentioned for content embedded within the image data itself.
Audit Metadata