omh-voice-input
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists of instructional content and metadata without any attached executable scripts, network requests, or direct system calls.- [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing untrusted voice-to-text data. It addresses this risk by mandating human confirmation gates for any irreversible or high-risk actions such as deletions, deployments, or transfers.- [SAFE]: Korean language strings in the routing signals section are verified as legitimate translations for voice-command terminology and do not exhibit characteristics of homoglyph obfuscation.
Audit Metadata