mixed-reality
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill implements data ingestion surfaces for external, untrusted inputs including QR code data and voice commands which could be manipulated to influence agent behavior.
- Ingestion points: Data from QR codes (
QRCodeTracker) and voice recognition keywords (VoiceCommandHandler) are ingested into the application context inreferences/examples.md. - Boundary markers: The provided code snippets do not include boundary markers or specific instructions to treat this external data as untrusted or to isolate it from the rest of the application logic.
- Capability inventory: As defined in
SKILL.md, the agent utilizing this skill has the capability to inspect the repository context, edit files, and execute build or test commands, creating a potential exploitation path if malicious data is processed. - Sanitization: The implementation examples do not demonstrate sanitization, filtering, or validation of the external strings before they are used to trigger actions or state changes.
Audit Metadata