p-video-avatar
Pass
Audited by Gen Agent Trust Hub on Aug 6, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is primarily instructional and uses a restricted set of tools defined in the frontmatter (Bash(belt *)). This limits the agent's shell access to a specific, legitimate CLI tool.\n- [EXTERNAL_DOWNLOADS]: The skill references installation scripts and documentation from github.com/inference-sh and inference.sh. These are well-known resources within the service's ecosystem and are documented neutrally.\n- [PROMPT_INJECTION]: While the disable_safety_filter parameter is mentioned, it is a configuration option for the external media generation model and does not target the agent's own behavioral constraints or safety protocols.\n- [DATA_EXFILTRATION]: Analysis confirms that no sensitive local files or environment variables are accessed. Network calls are exclusively associated with the service's operational requirements.\n- [COMMAND_EXECUTION]: Shell commands are used to invoke the belt CLI for remote application execution, which is the intended and transparent purpose of the skill.
Audit Metadata