p-video-avatar
Pass
Audited by Gen Agent Trust Hub on Aug 17, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill documentation identifies a
disable_safety_filterparameter that is enabled by default (true). This instructs the AI model to bypass content safety protocols, which could lead to the generation of prohibited or harmful content. - [EXTERNAL_DOWNLOADS]: The skill requires the installation of the
beltCLI and associated skills from theinference-shorganization. These resources are hosted on external domains (GitHub, NPM) that are not recognized as part of the trusted vendor list. - [COMMAND_EXECUTION]: The skill uses the
Bashtool to run thebeltcommand, which processes user-provided inputs (images and text) and interacts with a third-party API atinference.sh. This execution model involves sending local context to an external provider. - [DATA_EXFILTRATION]: User-provided portraits and voice scripts are ingested by the skill and transmitted to the external
inference.shservice for video generation, representing a data ingestion and transmission path to a third-party provider.
Audit Metadata